PSY 335 Week 3 Experimental and Quasi-Experimental Designs Example

Randomize, Cross Over or Compare Dorms? Choosing a Design to Test Phones and Sleep

[Student Name]

University of Phoenix

PSY/335: Research Methods

Week 3 Assignment

[Instructor Name]

[Date]

The student researcher, the residence halls and the designs are composites written for a model paper; research findings come from the sources listed.

An experiment is the strongest tool psychology has for testing cause and effect, but only when its design rules out rival explanations. This paper compares three designs for testing whether putting phones away before bed helps college students sleep, examines what each controls and leaves open and recommends one.

The Question and the Variables

Marisol, the composite psychology student from earlier weeks, now has an ethically reviewed plan. Her independent variable is evening phone use, with two levels: phone put away from one hour before lights out, or the student's usual routine. Her main dependent variable is sleep onset latency, the minutes between lights out and falling asleep, measured with a wrist-worn activity tracker. Secondary outcomes are total sleep time and morning alertness on a brief reaction-time test. Extraneous variables include caffeine, exams, work schedules and roommates.

Evidence From a Related Trial

He et al. (2020) randomly assigned university students in China who used phones heavily at bedtime to either restrict phone use for thirty minutes before sleep or continue as usual for four weeks. Students in the restriction group showed shorter sleep onset, better sleep quality, lower pre-sleep arousal, improved mood and better working memory. The trial was a pilot with a modest sample, but it gives Marisol both a model to follow and a rough idea of the effect sizes to expect, which matter for deciding how many participants she needs.

Design One: Randomized Two-Group Experiment

Sixty students are randomly assigned, thirty to the phone-away condition and thirty to their usual routine, for two weeks. Random assignment is the key strength: differences between people, such as night-owl tendencies, anxiety or heavy course loads, should spread evenly across groups, so a difference in sleep can be attributed to the condition rather than to who was in each group.

The weakness is variability. People differ widely in how quickly they fall asleep, and with thirty per group, that natural spread may hide a modest effect. Two other problems remain. Participants know which condition they are in, and the control group may change its habits after learning what the study is about.

Design Two: Crossover Within-Subjects Design

Every participant completes two one-week periods, one with the phone put away and one with the usual routine, separated by a few days. Half do the phone-away week first and half do it second. Because each person serves as their own comparison, stable individual differences drop out, and fewer participants are needed for the same precision.

The design opens other threats. Order effects arise if the first week changes the second, for example if students who give up the phone first keep the habit. Counterbalancing spreads such effects across conditions, and a short washout gap helps. History is another threat: if midterms fall in one week, sleep may worsen for reasons unrelated to phones. Scheduling both periods away from exam weeks reduces this risk.

Design Three: Quasi-Experiment Across Residence Halls

Suppose one residence hall launches a voluntary phone-free hour with a lights-down policy, and a second hall does not. Marisol could compare sleep in the two halls before and after the policy begins. This nonequivalent groups design with a pretest and posttest is practical, since the change happens anyway, and it observes behavior in students' real living spaces.

The cost is internal validity. Students were not randomly assigned to halls; one hall may house more first-year students, athletes or honors students, a selection threat. Anything else that changes in one hall at the same time, such as a noisy renovation, is a history threat that could mimic or mask an effect. The pretest helps by showing whether the halls started out similar, but it cannot rule out every difference.

Random assignment does what no amount of measuring can: it spreads every unmeasured difference between groups by chance.

Threats Shared by All Three

Shadish et al. (2002) described a set of threats to the validity of causal conclusions, including selection, history, maturation, testing, instrumentation and attrition, and explained how design features such as random assignment, pretests and comparison groups guard against them. Two threats apply to every design here.

Attrition is the first: students who find the phone-free hour hardest may drop out, leaving a group that looks better than it would have. Recording reasons for withdrawal and analyzing all participants as assigned addresses this.

Expectation is the second. Orne (1962) argued that participants actively look for cues about what a study is testing and often try to behave as good subjects, a pattern he called demand characteristics. Students who believe phones harm sleep may report sleeping better in the phone-away condition simply because they expect to. Using the activity tracker rather than self-report for the main outcome, and stating the purpose in general terms, reduces this risk.

External Validity

All three designs study volunteers at one university over a short period. Results may not generalize to students who did not volunteer, to older adults or to longer stretches of time. The residence hall design has the most realistic setting, while the randomized designs trade some realism for control.

Comparing the Designs

The randomized two-group design offers strong control but needs many participants to detect a modest effect. The crossover design keeps random ordering, cuts the influence of individual differences and needs fewer people, but requires care with order and timing. The residence hall design is realistic and convenient but leaves selection and history open.

Recommendation

The crossover design best fits Marisol's question and resources. It keeps the causal strength of an experiment, handles large individual differences in sleep and is fair, because every participant tries the phone-free routine. Counterbalanced order, a washout gap, scheduling away from exams, objective sleep measures and a general description of purpose address its main threats.

Conclusion

Each design answers the question with different strengths. Random assignment gives the experimental designs their power to support cause and effect, the crossover design adds efficiency and the quasi-experiment adds realism at the cost of control. Weighing threats against practical limits leads to a crossover experiment for Marisol's study.

References

He, J., Tu, Z., Xiao, L., Su, T., & Tang, Y. (2020). Effect of restricting bedtime mobile phone use on sleep, arousal, mood, and working memory: A randomized pilot trial. PLoS ONE, 15(2), Article e0228756. https://doi.org/10.1371/journal.pone.0228756

Orne, M. T. (1962). On the social psychology of the psychological experiment: With particular reference to demand characteristics and their implications. American Psychologist, 17(11), 776-783. https://doi.org/10.1037/h0043424

Shadish, W. R., Cook, T. D., & Campbell, D. T. (2002). Experimental and quasi-experimental designs for generalized causal inference. Houghton Mifflin.