C: Method example
The survey design and data for the example Method section below is adapted from a classroom project by Fettig, López Fuentes, and Villarreal (2019):164
Method
Participants
There were a total of 37 participants who completed the survey. Twenty-three of them were women (62.2%), and the remainder were men (n = 14, 37.8%). The ages (in years) ranged from 18 to 22 (M = 19.1, SD = 0.95), with the men being slightly older and more varied in age (M = 19.4, SD = 1.09) than the women (M = 19.0, SD = 0.825). The mean, the median, and the mode of age were all equal at 19.
The participants were recruited via convenience sampling through two GroupMe accounts shared by students from two large classes in Psychological & Brain Sciences at Texas A & M from the previous year.
Materials
The survey consisted of eleven questions overall. Two of the variables measured participant demographics: Sex (male vs. female); and Age (in years). Two other questions had to do with administration (covered in Procedures below). The remaining seven variables had directly to do with the research questions at hand. These are covered next in the order in which they appeared in the survey to the participants.
Outcome variables. There were three outcome variables that were designed to be averaged into a single composite, outcome variable: Sleep Quality. The first, Sleep Satisfaction, was worded as follows: “How satisfied are you with the amount of sleep you get each night?” Responses were measured on a Likert scale from 1 to 5, with 1 corresponding to “Not at all satisfied,” and 5 corresponding to “Very satisfied.” The second variable Restedness, was worded in the survey as follows: “On a scale of 1 to 5, how rested do you feel on average?” This question used the same scale, with 1 corresponding to “Not at all rested,” and 5 corresponding to “Very rested.” Finally, the third outcome variable, Typical Hours Slept, was simply the number of hours that the participant estimated they slept per night. This number was entered directly by the participant.
Predictor variables. There were four predictor variables. The first two were nominal variables. One was a two-level nominal variable called Extracurricular Activities.The question that corresponded to the variable was as follows: “Are you a part of extracurricular activities? i.e., clubs, sports, greek life.” Participants could answer either “Yes” or “No.” The other nominal variable was a three-level variable called Employment Status (“Are you employed?”). There were three options for the participant: “No,” “Full-time,” and “Part-time.” The next two predictor variables were continuous. The first of these was Academic Hours (“How many academic hours are you currently taking?”). Participants entered the number of credit hours they were taking that semester. The second (and last of the four predictor variables) was Workload Intensity: “How do you rate the intensity of your overall workload, including academic, work-related, and extracurricular commitments?” This was measured on a 1-5 Likert scale with 1 corresponding to “Not intense at all,” and 5 corresponding to “Very intense.”
Design & Analysis
We were interested in how Sleep Quality (the composite, outcome variable) was predicted by whether students were involved in extracurricular activities, the degree to which they were employed, their current academic course load, and the intensity of their general workload (work, academic, and otherwise).
The calculation of the composite variable, Sleep Quality, first required the conversion of the three, raw outcome variables into z-scores, and from there into a composite score. Thus, the following variables were converted to z-scores: Sleep Satisfaction (M = 3, SD = 1); Restedness (M = 2.92, SD = 1.12); and Typical Hours Slept (M = 6.45, SD = 1.12).
After the standardization into z-scores, the three newly standardized scores were averaged into the single composite. An internal reliability analysis showed that item consistency was quite high (Cronbach’s Alpha = 0.86). An analysis at the Alpha-level change due to single-item dropping revealed that removing any of the items would result in a drop in Alpha. Therefore, all three variables were retained for the composite. The resulting z-scored composite, Sleep Quality (M = 0, SD = 0.883) ranged from a score of -2.39 to a maximum score of 1.73. This served as the single outcome variable in all subsequent analyses.
Summaries of the predictor variables follow. Six of the participants reported engaging in no extracurricular activities (Extracurricular Activities) whatsoever, whereas 31 reported engaging in one or more such activities. We would expect extracurricular activities to interfere with sleep, so those participating in no extracurricular activities should report higher levels of Sleep Quality.
With respect to employment status (Employment Status), one participant reported working full-time, eight reported working part-time, and 28 reported not working at all. Work can interfere with academics, so one would expect that those working part time should express less sleep quality than those not working at all.
The participants reported being markedly more overworked (Workload Intensity) than the neutral, middle part, 3, on the 1-5 scale (M = 3.92, SD = 0.76, with scores ranging from 3 to 5, meaning that no one reported being more rested than usual). Scores on Workload Intensity should correlate negatively with Sleep Quality (as scores go up on one, they go down on the other).
They also reported taking an average of 14.4 credit hours (Academic Hours) for the semester (SD = 1.59), with values ranging from 10 to 17. This variable should associate negatively with Sleep Quality since the more classes one is taking, the more one has to study, presumably, leaving less time for sleep.
Procedures
The researchers simply sent the link to this survey out to potential respondents in the GroupMe list mentioned above under Participants. Upon receiving the survey, participants were informed about the nature of the study, that participation was optional, and that no personally identifying information would be recorded. At this point, there were given the opportunity to opt out of the study.
If they continued with the study, they were asked all the questions mentioned above in the same order as above. Importantly, the participants were not required to respond to any of the questions; they could leave any response blank if they so chose. Finally, at the end of the study, participants were given the option to exclude their data from analysis. None of the participants chose this option.
After two weeks of soliciting responses, the final data set was selected for analysis.
References
Permission to use their data was given each author separately in late September, 2019, via a series of email exchanges.↩︎