22
Time of Participation 1 Consistency in the Time of Experiment Participation and Personality Correlates: A Methodological Note John M. Zelenski* Washington University Cheryl L. Rusting State University of New York at Buffalo Randy J. Larsen Washington University Zelenski, J. M., Rusting, C. L., & Larsen, R. J. (2003). Consistency in the time of experiment participation and personality correlates: A methodological note. Personality and Individual Differences, 34(4), 547-558. *John M. Zelenski now at Carleton University in Ottawa, Canada John M. Zelenski Department of Psychology Carleton University 1125 Colonel By Drive Ottawa, ON K1S 5B6 Canada phone: (613) 520-2600 ext. 1609 fax: (613) 520-3667 email: [email protected] web: http://www.carleton.ca/~jzelensk/

Consistency in the time of experiment participation and personality correlates: a methodological note

Embed Size (px)

Citation preview

Time of Participation 1

Consistency in the Time of Experiment Participation and Personality Correlates:

A Methodological Note

John M. Zelenski*

Washington University

Cheryl L. Rusting

State University of New York at Buffalo

Randy J. Larsen

Washington University

Zelenski, J. M., Rusting, C. L., & Larsen, R. J. (2003). Consistency in the time of

experiment participation and personality correlates: A methodological note.

Personality and Individual Differences, 34(4), 547-558.

*John M. Zelenski now at Carleton University in Ottawa, Canada

John M. Zelenski Department of Psychology Carleton University 1125 Colonel By Drive Ottawa, ON K1S 5B6 Canada phone: (613) 520-2600 ext. 1609 fax: (613) 520-3667 email: [email protected] web: http://www.carleton.ca/~jzelensk/

Time of Participation 2

Abstract

Certain voluntary human participant pool procedures (e.g., where the participant

selects when to participate in experiments) can compromise the representativeness of

samples and the efficacy of random assignment procedures. In this study we recorded the

time and date that university students participated in two experiments run during two

different months in the same semester, and found significant test-retest correlations

between times of participation. These correlations suggest that when students participate

in psychological research may be a stable characteristic related to personality. We then

correlated composite (combined across the two experiments) time and date scores with an

array of standard demographic and personality measures, and found that experiment

participation time was related to several personality variables (e.g., sensation seeking,

morningness, achievement motivation). We conclude that, although the effect is likely to

be small, some voluntary sign-up procedures and experimental procedures can introduce

systematic bias in the form of personality differences.

Keywords: student participants, subject pool, personality, consistency, experimental

design, time of participation, time of day, validity

Time of Participation 3

Consistency in the Time of Experiment Participation and Personality Correlates:

A Methodological Note

The vast majority of American psychological research makes use of

undergraduate student participant pools. Typically, students in lower level psychology

classes fulfill course requirements or gain extra credit by participating in experiments.

Such student pools have provided psychologists with a convenient means to recruit

participants. In most cases, researchers simply advertise an experiment, and participants

choose a time that is agreeable to them. Although such procedures are convenient for

both researchers and participants, their voluntary nature raises an important question: do

different kinds of people participate in psychological experiments at different times? That

is, do stable individual differences contribute to when students choose to participate in

experiments?

This question can be asked for time of day, e.g., Do morning types tend to

participate in experiments earlier in the day and evening types later?. Also, we can ask

the question about time of the semester, e.g., Do different types of people come at

different times of the semester?. It is possible that the students who sign up for times in

the first week of the semester differ from those who put off their experimental

participation to the end of the semester. Intuitively, differences between early and late

participants seem likely.

More than a simple curiosity, individual differences in participation times can

have potentially important consequences for the representativeness of samples and

random assignment procedures. Most voluntary participant pool procedures and

experiment methodologies allow for participation times that are not equally distributed

across a semester or times of day. If an experiment is conducted during only the first half

of a semester, or only in the mornings, for example, the sample may not be representative

of even the minimal diversity found in college participant pools. An even more troubling

Time of Participation 4

situation could arise if participation times systematically covaried with assignment to

different experimental conditions. That is, assignment to conditions is not truly random if

it varies over time (and thus also personality). Therefore, we must be concerned about the

possibility that individual differences may be related to participation times.

Previous research has focused on time-related differences in participants'

performance on various laboratory tasks. These types of studies have yielded mixed

evidence for differences between participants who participate early versus late in a

semester. For example, Underwood, Schwenn, and Keppel (1964) failed to find

significant differences between early and late participants across four semesters of

experiments using a paired-associates learning task. Langston, Ohnesorge, Kruley, and

Haase (1994) failed to find statistically significant differences between early and late

participants in both text comprehension and signal detection tasks. Moreover, Langston

et. al. (1994) performed power analyses that suggested their methods had sufficient

power to detect potential differences. Conversely, Richter, Wilson, Milner, and Senter

(1981) did find statistically significant differences on two different types of tasks, serial

learning and symbol substitution. However, the pattern of results differed by task. Early

semester participants performed better on the serial learning task, whereas late

participants performed better on the symbol substitution task. Given these mixed

findings, explanations for why or how early and late participants differ remain

mysterious.

Wang and Jentsch (1998) conducted a study that compared early and late

participants' performance on a paired-associates learning task, but unlike the previously

discussed experiments, they also measured personality traits. Like Underwood et. al.

(1964), they found no significant differences between early and late semester participants

on the paired-associates task. However, they did find differences on some of the

personality measures. Early participants had a more internal academic locus of control,

higher levels of work orientation, and lower levels of competitiveness than late

Time of Participation 5

participants (but null findings were obtained on other scales such as need for cognition

and procrastination). Wang and Jentsch (1998) concluded that although motivation and

personality may differ between early and late participants, these differences do not affect

performance-based variables, such as paired-associates learning.

Although, this conclusion does not account for Richter et al. 's (1981) findings, it

does seem consistent with the results of other studies. That is, studies that have included

motivation and personality measures have generally been more successful in

demonstrating differences between early and late participants. Much of the personality

and motivation based research can be summed up by what might be called a 'good

student' effect. Early participants have tended to be more socially responsible as indicated

by Personality Research Form profiles (Holden & Reddon, 1987), more intrinsically

motivated (Hom, 1987), more compliant as indicated by oral Rorschach imagery

(Masling, O'Neill, & Jayne, 1981), more curious and conforming (Roman, Moskowitz,

Stein, & Eisenberg, 1995), more academically and achievement oriented, and have

achieved higher standardized test scores (Evans & Donnerstien, 1974). Additionally,

early term participants tend to score higher on measures of Type A personality (Strube,

1982), personal need for structure (Neuberg & Newsom, 1993; Roman et. al., 1995),

introversion (Liberty, 1993), and internal locus of control (Evans & Donnerstein, 1974).

Female participants have also tended to participate earlier than males (Richter et. al.,

1981). Cooper, Baumgardner, and Strathman (1991) reported that early participants were

more likely to be female, freshmen, Protestant, from small towns, and low in self-

consciousness. However, these findings were based on a large sample, and thus the rather

small effects were statistically, but perhaps not practically, significant ('significant'

correlations were as low as .07).

In summary, research has demonstrated some significant (if small) differences

between people who participate early as opposed to later in a semester in performance on

laboratory tasks, motivation, and personality variables. However, almost no research has

Time of Participation 6

addressed possible time of day differences between research participants. (The Cooper et.

al. (1991) study is an exception, but found only one significant difference.) Thus, the

question of whether or not individual differences influence the time of day a student

participates in psychological research remains open. Our study attempts to remedy this

problem by investigating personality differences with respect to time of day differences

in participation, in addition to investigating time of semester differences.

Past attempts to demonstrate individual differences in participation times have

usually followed the same general procedure. Researchers have recorded when their

participants participated in an experiment, and then made comparisons on some

dependent variable (e.g., a paired-associates task or personality questionnaire). In other

words, measurement of the individual difference in question (participation time) has

relied on only one observation per participant. Single instances of behavior are usually

poor indicators of stable individual differences (Mischel, 1968; Diener & Larsen, 1984).

The signal to noise ratio (i.e., trait to random factors) of single measurements is usually

too low to obtain reliable estimates of a trait's strength. This may account for some of the

inconsistency of past findings. Additionally, since past research has only measured

participation time once, it has not been able to address the question of whether or not

participation time can be considered a stable and consistent individual difference. That is,

no one has demonstrated, or even tested, the 'traitness' of participation time.

In the present study we address this issue by measuring participation time in the

same participants across two separate experiments. This allows us to measure

consistency, or, in other words, to calculate a correlation between the participation times

of two different experiments (i.e., a test-retest correlation). Assuming that some degree of

consistency can be demonstrated, the two measures of participation time can be

combined to create a composite score. Moreover, this composite score should be a more

reliable indicator of each participant's tendency to participate in experiments early versus

late in the semester and in the morning versus afternoon/evening. Although two instances

Time of Participation 7

of behavior are far from ideal, the combined score should be more reliable than a single

observation. Using a more reliable measure of participation time should increase the

chances of finding differences between early and late participants.

To summarize, we have two important objectives for this study. First, we attempt

to demonstrate consistency in participation time for both time of day and for time over

the course of the experiment. Second, we attempt to predict this tendency (early versus

late participation) with more standard personality and demographic variables. To these

ends, we measured participation time over two experiments, and measured personality

with a variety of questionnaires (described below). Our study is unique because we are

among the first to measure participation time over multiple occasions. Moreover, almost

no previous research has considered possible personality related time of day differences

in experiment participation.

We selected personality variables using three criteria. First, we considered

personality scales and demographic characteristics suggested by past research (e.g., type

A, birth order, and sex). Second, we chose two inventories that purport to measure the

major characteristics of human personality, the NEO-FFI (Costa & McCrae, 1992) and

the ZKPQ, a measure of Zuckerman's 'Alternative Big Five' (Zuckerman, Kuhlman,

Thornquist, & Kiers, 1991). Finally, we included two questionnaires that theoretically

should be related to the time of experiment participation. These were the Morningness-

Eveningness Questionnaire (MEQ, Horne & Östberg, 1976), and the Sensation Seeking

Scale (Zuckerman, 1971), There is a clear prediction that morningness will correlate with

participation earlier in the day. Assuming that sensation seekers do not find experiment

participation very thrilling, we predict that sensation seekers will participate later in the

experimental period.

Method

Participants. All participants were enrolled in an experiential laboratory course in

psychological research. There were no constraints on who could enroll in this course, and

Time of Participation 8

the course was advertised widely across campus. (Nonetheless, about 75% of the students

were psychology majors.) As part of the course's requirements, students filled out

numerous personality questionnaires and participated in two laboratory experiments.

There were 86 students enrolled in the class, but only 80 of the students participated in

both experiments. Thus, the sample reported here consisted of 20 men and 60 women

with a mean age of 20.46 years. Students were not made aware of this study's purpose or

hypotheses.

Materials. Demographic characteristics were collected with a questionnaire specially

designed for this study. Additionally, the following widely used questionnaires provided

scores on personality variables.

The Jenkins Activity Survey (Jenkins, Rosenman, & Zyzanski, 1974) is a 43-item

self-report questionnaire that provides a measure of Type A personality. Items are

multiple choice questions with two to five options (i.e., different questions have different

numbers of options). Despite the misleading name ('type'), scores on this questionnaire

are dimensional rather than categorical. Type A behavior pattern is characterized by time

urgency, hostility, and concern with achievement.

The Locus of Control Scale (Rotter, 1971) is a 40-item yes/no questionnaire that

assesses the degree to which a person has an internal (as opposed to external) locus of

control. Locus of control refers to an individual's beliefs about causal relationships. For

example, a person with an external locus of control thinks they have little personal

control over their life, and instead tends to attribute causality to external sources.

The Personality Research Form (Jackson, 1967) short form contains 80 true/false

items. Of interest to our study were the two 16-item scales of need for achievement and

need for power. The PRF was designed to measure Murray's (1938) concept of

psychogenic needs or motives (c.f., McClelland, Koestner, & Weinberger, 1989).

The Adjective Check List (Gough & Heilbrun, 1965) consists of 300 potentially

self-descriptive adjectives that are either endorsed or not endorsed by the test taker. Of

Time of Participation 9

interest to this study were scales designed to assess need for achievement (38 items) and

need for dominance (40 items). As with the PRF, these scales were designed to measure

Murray's (1938) needs.

The Crowne-Marlowe Scale (Crowne & Marlowe, 1964) contains 33 true/false

items, and assesses the trait of social desirability. Social desirability refers to the

tendency to respond to questionnaire items in a socially acceptable manner. Such socially

desirable responses may be motivated rather than an accurate representation of 'truth'.

The NEO-Five-Factor Inventory (Costa & McCrae, 1992) contains 60 potentially

self-descriptive items which are rated on a 5-point scale that ranges from 'strongly

disagree' to 'strongly agree'. The NEO-FFI has five 12-item scales designed to measure

the 'Big 5' traits of neuroticism, extraversion, openness, agreeableness, and

conscientiousness (c.f., Saucier, 1998). See John (1990) for further information on the

Five-Factor model of personality.

The Zuckerman-Kuhlman Personality Questionnaire (Zuckerman, et al., 1991)

contains 99 true/false questions. The ZKPQ has five factor analytically derived

personality scales designed to measure an 'alternative Big 5' set of traits: impulsive-

sensation seeking (19 items), sociability (17 items), neuroticism-anxiety (19 items),

hostility-aggressiveness (17 items), and activity level (17 items). The ZKPQ was

designed to assess biologically based dimensions of personality. See Zuckerman et al.

(1991) for more information on the 'Alternative Big 5' model.

The Morningness-Eveningness Questionnaire (Horne & Östberg, 1976) contains

19 items, some of which are multiple choice, and others that ask about preferred times for

activities using a 24 hour day as a response scale. The MEQ contains a single scale

designed to assess circadian rhythms in terms of 'feeling best' times. For example, people

scoring high on morningness report rising earlier and feeling more alert in the morning

than people scoring low on morningness.

Time of Participation 10

The Sensation Seeking Scale (Zuckerman, 1971) contains 72 forced choice items

with two self-descriptive options per item. Sensation seeking refers to a preference for

novel, impulsive, and thrilling activities.

Procedure. Personality questionnaires were distributed to students during a class period,

and completed either during class time or at home (i.e., some questionnaires were

assigned as homework). Questionnaires completed at home were returned at the next

class meeting one week later. The students were also recruited for two laboratory

experiments on personality and emotion. (Further details can be found in Rusting (1999)

for the first experiment, and Zelenski and Larsen (1999) for the second experiment.) All

students in the sample were required to participate in these two experiments, or complete

an alternative assignment (i.e., other than experiment participation). The sign-up

procedure for both experiments was nearly identical. During the class period before the

experiment would begin, a sign-up sheet was distributed in class. This sign-up sheet

listed time slots for one month. Although most students chose their participation times

during this first distribution of the sign-up sheet, the sheet was distributed each week in

class until the experiment was complete. A few students scheduled their participation by

contacting one of the experimenters by phone or electronic mail. The first of the two

experiments was conducted in February, and the second during the last week of March

and the first three weeks of April. Both experiments ran every day of the week except

Sunday, and almost every time of day (between 9 AM and 8 PM). Time of participation

was recorded when each student actually arrived at the laboratory (i.e., data are not based

on the sign-up sheets).

Results

Descriptive statistics for all questionnaire measures are reported in Table 1. The

first experiment was run across 24 days. The median day of participation was day number

15, with a standard deviation of 8.07 days. The mean time of participation was 1:34 PM,

with a standard deviation of 2.56 hours. The second experiment was run across 30 days.

Time of Participation 11

The median day of participation was day number 18, with a standard deviation of 9.02

days. The mean time of participation was 1:55 PM, with a standard deviation of 2.25

hours.

To assess the degree of consistency in the time of experimental participation, we

correlated participation times across the two experiments. For time of day, r = .39, p <

.001. For day of participation, r = .38, p < .001. Thus, there is a moderate degree of

consistency in when these students participated in the two experiments, both for day

number and time of day. Because we found consistency, we combined data across both

experiments. Time and day composite scores (i.e., sums of day numbers and military

times) for the two experiments were correlated with demographic and personality

variables, and these correlations are presented in Table 2.

Table 2 reveals statistically significant correlations between personality variables

and the time over the course of the experiments when students participated (i.e., day

number). As predicted, the sensation seeking scale correlated with day number such that

participants scoring higher on sensation seeking participated later. Similarly, the

impulsive-sensation seeking scale of the ZKPQ was significantly correlated with later

participation. Morningness also correlated significantly with day number, such that

morning types tended to participate earlier in the experimental period. Table 2 indicates

marginally significant (i.e., p < .10) correlations between day number and birth order,

need for power, and neuroticism. Finally, a t-test suggested a marginally significant trend

such that females participated earlier than males (t = 1.82, p = .07)

The right hand column of Table 2 shows that personality measures were also

correlated with the time of day students participated in experiments. As predicted,

morningness was significantly correlated with participation earlier in the day. The

significant negative correlations of time of day with conscientiousness, social desirability,

and need for achievement indicate that students scoring higher on these variables tended

to participate earlier in the day. Activity level and extraversion also correlated negatively

Time of Participation 12

with time of day at marginally significant levels (p < .10). Birth order correlated with

time of day such that early-borns tended to participate earlier in the day than later-borns.

Finally, a t-test suggests that females participated earlier in the day than males (t = 3.13, p

< .01).

Discussion

A major goal of this study was to test for consistency in when students participate

in psychological experiments. That is, to answer the question of whether or not the self-

selection of experiment participation time appears trait-like, i.e., it is stable over time and

across situations. The results of this study suggest that there is indeed a modest degree of

consistency in the time of experiment participation, both for time of day, and time over

the course of the experiment. We base this conclusion on the fact that, across two

experiments, time of participation was correlated. Moreover, these tendencies, to arrive

early or late for experiment participation (time and day), were successfully predicted by

standard personality and demographic measures. In other words, time of participation

seems to covary with meaningful individual differences.

The results of this study replicate findings that males and those high in the need

for power tend to participate later in the experimental period. Holden and Reddon (1987)

interpreted the finding that students high in need for dominance participate later as due to

a lower sense of social responsibility. In terms of experiment participation, they probably

feel less compelled to fulfil their participation obligations, and thus, put it off longer.

Males may participate later because they are less interested in the experiments or

psychology in general. (Consider that psychology majors are more likely to be female.)

Despite replicating findings for gender and need for dominance, previously reported

correlations between time of participation over the course of an experiment and Type A

personality, locus of control, year in college, and achievement motivation were not

replicated in our study. The low internal consistency observed in this sample on the type

Time of Participation 13

A and locus of control scales may have contributed to null findings. Alternatively, these

failures to replicate may be due to differences between our procedure and the procedures

used by other researchers. For example, we measured participation time over one month

(twice), whereas most other studies considered participation time over an entire semester.

We also measured some personality variables that have not been included in

previous research. For example, we found that those scoring high in impulsive-sensation

seeking tended to participate later. It seems unsurprising that these students would put off

an activity that they would probably find mundane, such as experiment participation.

Somewhat surprisingly, we also found that morningness correlated with earlier

participation over the course of the experiments. Perhaps morningness relates to longer

rhythms or a more general tendency to do things earlier.

This study also examined time of day differences in experiment participation.

Almost no previous research has addressed this issue, so our findings can be seen as a

first step towards determining which, if any, personality characteristics covary with time

of day differences in experiment participation. As expected, we found that participants

scoring higher on morningness were more likely to participate early in the day. Later-

borns and those scoring low in need for achievement and conscientiousness also

participated later in the day. These three findings come together nicely to suggest that

students concerned with work and achievement participate in experiments earlier in the

day. Males and those scoring low in social desirability and activity level tended to

participate later in the day, but it is difficult to make theoretical explanations for these

findings. We replicated Cooper et al.'s (1991) finding that extraverts tend to participate

earlier in the day. This finding was surprising because extraversion often correlates with

eveningness (Larsen, 1985, Neubauer, 1992). Perhaps extraverts participate in

experiments early in the day in order to 'get it out of the way'.

In summary, personality successfully predicted time of experiment participation.

Given previous findings and coherence across different measures in this study, it seems

Time of Participation 14

likely that gender, achievement motivation, and impulsivity are related to when students

participate in experiments. Of course, further replication in future studies would

strengthen these suggestions.

Disregarding standard personality variables, self-selection of experiment

participation time appears to be a trait-like characteristic in that it was stable over two

experiments. When compared with test-retest correlations typically found for personality

questionnaires, the correlations of .38 and .39 obtained here may seem small. However, a

more appropriate comparison might be the inter-item correlations for a questionnaire,

which are typically much smaller (often around .30). The consistency correlations

reported here describe the relationship between two single items of behavior. With only

two instances of behavior, random (as opposed to trait-related) factors can easily reduce

the correlation. Thus, we feel that correlations of .38 and .39 constitute an important

demonstration of consistency. Future research could attempt to gather participation times

across more experiments, and thus further demonstrate the 'traitness' of this behavior. For

example, if participation times were collected across five experiments (with the same

participants), it would be reasonable to calculate a reliability estimate (alpha) as if each

experiment was a different question on a scale designed to measure participation time.

The Spearman-Brown formula suggests that increasing the number of observations could

substantially increase the reliability of such a participation time measure. These more

reliable estimates of participation time could also be used to better identify which

standard personality scales best predict time of participation.

On the other hand, the correlation between just two experiments is useful because

it estimates the size of personality's effect on experiment participation time for a single

experiment. In other words, the correlation between two experiments' times estimates the

variance that personality can account for in when students participate in a single

experiment. The variance shared between the participation times of two experiments

includes the effect of all personality differences related to participation time. Our results

Time of Participation 15

suggest that if we were able to measure personality completely and predict the

participation time for one experiment, we could predict about 15% of the variance. So,

even if participation time aggregated over many experiments turns out to be very

consistent, it seems likely that personality can have only a small to modest effect on the

participation times of any one experiment. This, of course, is an empirical question.

Because we are the first to report such a correlation, it is possible that future research

may demonstrate significantly larger relationships. In any case, even a small to modest

effect can have important consequences, so it should not be dismissed.

The findings that time of experiment participation is consistent and related to

personality have clear implications for laboratory based psychological research. Most

obviously, experiments that recruit participants from student pools at only one time of

day or at one time of the semester compromise the representativeness of their samples. In

personality research, such practices will limit the variability in certain personality

characteristics, and thus potentially decrease the probability of finding significant results.

Of course experimental research should also be concerned with recruiting representative

samples to improve the generalizability of findings. Even if participants are recruited at

many times across an entire semester, problems can still arise. For example, the

experimenters (lab personnel) may change over the course of an experiment, or

experimenters may run participants only at one time of day. In such cases, personality

becomes correlated with experimenters and any differences in the way they implement

the experimental protocol. Likewise, personality will be correlated with experimenter

drift over time in how the protocol is carried out (Strube, 1982).

In a worst case scenario, personality differences related to time of participation

could be correlated with assignment to treatment conditions. For example, a researcher

may decide post hoc to run an additional condition at the end of a semester. Such a

decision could create spurious results due to personality. Consider the following example:

trait A correlates .40 with a dependent variable of interest (e.g., a reaction time), and trait

Time of Participation 16

A also correlates .40 with assignment to a treatment condition (due to time of

participation personality differences). Let us further assume there is no effect of the

treatment condition. Using the chain rule from path analysis, we can estimate a

completely spurious effect of the treatment condition equal to .16. Fortunately, this type

of error is completely avoidable by following the principle of random assignment to

conditions. The results of our study suggest that researchers adopt methods that maximize

representativeness, and do not allow for such spurious findings.

In all fairness, the potential for time-related personality differences to influence

the results of laboratory experiments is limited. In the example above, we assumed a

relatively strong (compared to our findings) correlation between personality and

assignment to condition. Moreover, personality will not always, or perhaps even not

often, correlate with the dependent variables of interest. Even when such correlations

exist, the effect sizes are likely to be small. Still, the potential for bias, even if remote,

should be an important concern for psychological researchers. Additionally,

representativeness may depend on when experimental sessions are conducted. We believe

we have presented enough evidence here to make time of experimentation an important

design consideration.

Time of Participation 17

References

Cooper, H., Baumgardner, A. H., & Strathman, A. (1991). Do students with

different characteristics take part in psychology experiments at different times of the

semester?. Journal of Personality, 59, 109-127.

Costa, P. T., & McCrae, R. R. (1992). Revised NEO Personality Inventory (NEO-

PI-R) and NEO Five-Factor Inventory (NEO-FFI) professional manual. Odessa, FL:

Psychological Assessment Resources.

Crowne, D. P., & Marlowe, D. (1964). The approval motive. New York: Wiley.

Diener, E. & Larsen, R. J. (1984). Temporal stability and cross-situational

consistency of affective, behavioral, and cognitive responses. Journal of Personality and

Social Psychology, 47, 871-883.

Evans, R. & Donnerstein, E. (1974). Some implications for psychological

research of early versus late term participation by college subjects. Journal of Research in

Personality, 8, 102-109.

Gough, H. G. & Heilbrun, A. B. (1965). The Adjective Check List manual. Palo

Alto, CA: Consulting Psychologists Press.

Holden, R. R. & Reddon, J. R. (1987). Temporal personality variations among

participants from a university subject pool. Psychological Reports, 60, 1247-1254.

Hom, H. L. Jr. (1987). A methodological note: Time of participation effects on

intrinsic motivation. Personality and Social Psychology Bulletin, 13, 210-215.

Horne, J. A. & Östberg, O. (1976). A self assessment questionnaire to determine

morningness-eveningness in human circadian rhythms. International Journal of

Chronobiology, 4, 97-110.

Jackson, D. N. (1967). Personality Research Form manual. Goshen, NY: Research

Psychologists Press.

Time of Participation 18

Jenkins, C. D., Rosenman, R. H, & Zyzanski, S. J. (1974). Prediction of clinical

coronary heart disease by a test for the coronary-prone behavior pattern, New England

Journal of Medicine, 290,1271-1275.

John, O. P. (1990). The "Big Five" factor taxonomy: Dimensions of personality in

the natural language and in questionnaires. In L. A. Pervin (Ed.), Handbook of

Personality: Theory and Research (pp. 66-100). New York: Guilford.

Langston, W., Ohnesorge, C., Kruley, P., & Haase, S. J. (1994). Changes in

subject performance during the semester: An empirical investigation. Psychonomic

Bulletin & Review, 1, 258-263.

Larsen, R. J. (1985). Individual differences in circadian activity rhythm and

personality. Personality and Individual Differences, 6, 305-311.

Liberty, H. J. (1993). The relationship between extraversion and time of data

collection. Personality and Individual Differences, 14, 835-836.

Masling, J., O'Neill, R., & Jayne, C. (1981). Orality and latency of volunteering to

serve as experimental subjects. Journal of Personality Assessment, 45, 20-22.

McClelland, D. C., Koestner, R. & Weinberger, J. (1989). How do self-attributed

and implicit motives differ?. Psychological Review, 96, 690-702.

Mischel, W. (1968). Personality and Assessment. New York: Wiley.

Murray, H. A. (with staff) (1938). Explorations in Personality. New York: Oxford

University Press.

Neubauer, A. C. (1992). Psychometric comparison of two circadian rhythm

questionnaires and their relationship with personality. Personality and Individual

Differences, 13, 125-131.

Neuberg, S. L. & Newsom, J. T. (1993). Personal need for structure: Individual

differences in the desire for simple structure. Journal of Personality and Social

Psychology, 65, 113-131.

Time of Participation 19

Richter, D. O., Wilson, S. D., Milner, M., & Senter, J. J. (1981). Some differences

among students volunteering as research subjects. Bulletin of the Psychonomic Society,

17, 261-263.

Roman, R. J., Moskowitz, G. B., Stein, M. I., & Eisenberg, R. F. (1995).

Individual differences in experiment participation: Structure, autonomy, and the time of

the semester. Journal of Personality, 63, 113-138.

Rotter, J. (1971). External control and internal control, Psychology Today, 5, 37-

42, 58-59.

Rusting, C. L. (1999). Interactive effects of personality and mood on emotion-

congruent memory and judgment. Journal of Personality and Social Psychology, 77,

1073-1086.

Saucier, G. (1998). Replicable item-cluster subcomponents in the NEO Five-

Factor Inventory. Journal of Personality Assessment, 70, 263-276.

Strube, M. J. (1982). Time urgency and Type A Behavior: A methodological note.

Personality and Social Psychology Bulletin, 8, 563-565.

Underwood, B. J., Schwenn, E., & Keppel, G. (1964). Verbal learning as related

to point of time in the school term. Journal of Verbal Learning and Verbal Behavior, 3,

222-225.

Wang, A. Y. & Jentsch, F. G. (1998). Point-of-time effects across the semester: Is

there a sampling bias?. The Journal of Psychology, 132, 211-219.

Zelenski, J. M. & Larsen, R. J. (1999). Susceptibility to affect: A comparison of

three personality taxonomies. Journal of Personality, 67, 761-791.

Zuckerman, M. (1971). Dimensions of sensation seeking, Journal of Consulting

and Clinical Psychology, 36, 45-52.

Zuckerman, M., Kuhlman, D. M., Thornquist, M., & Kiers, H. (1991). Five (or

three) robust questionnaire scale factors of personality without culture. Personality and

Individual Differences, 12, 929-941.

Time of Participation 20

Author Note

Portions of this paper were presented at the 2000 meeting of the Society for

Personality and Social Psychology in Nashville, TN. We would like to thank Michael

Strube for his helpful comments. Correspondence regarding this article should be

addressed to John M. Zelenski, Department of Psychology, Washington University,

Campus Box 1125, St. Louis, MO 63130, U.S.A., or sent by electronic mail to

[email protected].

Time of Participation 21

Table 1. Descriptive Statistics Predictor Mean SD Range Alpha Semester in College 6.01 1.68 0-8 -- Credits this Semester 15.12 1.83 10-19 -- Birth Order 1.72 1.02 1-6 -- Type A 101.81 8.62 81-116 .58 Internal Locus of Control .24 .10 .03-.50 .60 Need for Achievement (A) 22.92 5.97 8-34 .83 Need for Achievement (P) 9.67 3.39 3-15 .72 Need for Dominance (A) 26.41 5.61 10-36 .78 Need for Power (P) 9.82 3.87 1-15 .83 Social Desirability 28.92 4.62 19-37 .33 Neuroticism 2.72 .59 1.33-4.00 .82 Extraversion 3.46 .52 2.08-4.67 .82 Openness 3.44 .50 2.25-4.50 .72 Agreeableness 3.68 .46 2.58-4.58 .76 Conscientiousness 3.52 .65 2.00-4.75 .89 Impulsive-SS 8.98 4.59 1-17 .84 Sociability 9.36 4.02 1-17 .81 Neuroticism-Anxiety 7.73 4.08 0-17 .80 Hostility-Aggressiveness 7.35 3.49 1-15 .75 Activity Level 7.65 4.00 0-16 .82 Morningness 40.54 9.29 23-67 .85 Sensation Seeking 34.74 9.78 12-71 .82

Time of Participation 22

Table 2. Correlations between personality and time of participation

Predictor Day of Experiment Time of Day Semester in College -.07 -.07 Credits this Semester -.08 .02 Birth Order1 .19* .23** Type A -.02 -.09 Internal Locus of Control -.05 .01 Need for Achievement (A) -.01 -.26** Need for Achievement (P) .04 -.18* Need for Dominance (A) .11 -.09 Need for Power (P)

.18* .10

Social Desirability .03 -.23** Neuroticism -.18* .09 Extraversion .12 -.19* Openness -.01 -.01 Agreeableness -.03 -.11 Conscientiousness -.17 -.23** Impulsive-SS .27** .03 Sociability .09 -.06 Neuroticism-Anxiety -.08 .06 Hostility-Aggressiveness .10 -.06 Activity Level .01 -.18* Morningness -.28** -.23** Sensation Seeking .22** .04

Note: 1 Only children were considered firstborns. *p<.10, **p<.05,