If you share a bed with your sweetheart, you might be interested in this example of correlational research summarized by PsyPost.  Their headline reads, “Study finds couples who cuddle at bedtime feel more secure and less stressed.”  

Here are some introductory statements:

Study authors Josh R. Novak and Kaleigh C. Miller noted that while couples often maintain their “own side” of the bed, they regularly cross this boundary for cuddling, affection, and sex. They also pointed out that people may adopt different sleep positions when sharing a bed compared to when sleeping alone.

The researchers … explored whether physical closeness during sleep was related to sleep disorders, perceived stress, attachment styles, and other relationship characteristics.

Here are some details about the study’s method. As you read, attend to the variables that were included in the study:

The study surveyed 143 heterosexual couples, comprising 286 individuals. The average age was 43 for men and 40 for women, and couples had been together for an average of 13 years. Fifty-two couples reported having the same chronotype (morning or evening preference).

Participants reported the degree of physical closeness in their typical sleep position at sleep onset. They also completed assessments of perceived stress, attachment insecurity (using the Experiences in Close Relationships Scale–Short Form), sleep disturbance (via the PROMIS Sleep Disturbance scale), daytime sleepiness, and their preferred individual sleeping position. Additional demographic and contextual information was collected, including relationship length, income, and whether children or pets shared the bed.

Questions:

1. What kind of claim is it to say “couples who cuddle at bedtime feel more secure and less stressed”–Frequency, association, or causal? 
2. Characterize the variables in the study, using this table. There are at least 10 variables in the description above, so focus on finding at least five of them. Be sure that you’ve located the key variables in the headline, though!

Variable name (Stated at the conceptual level)	Potential levels of this variable (if you don’t know the numbers, you can say “low to high” or maybe “yes or no”)	Is this variable manipulated or measured?

3. This is a correlational study. Why? 

Here’s the result: 

…couples who reported physically closer sleep positions—such as spooning, sleeping intertwined, or face-to-face—also reported lower levels of perceived stress and lower levels of anxious or avoidant attachment.

1. Sketch a graph of the result for sleep position and stress. Will you choose a bar graph or a scatterplot? If you classified “sleep position” as a categorical variable, you’ll probably make a bar graph, but make sure you have all possible levels represented (that is, you’ll need some non-cuddling levels as well). If you see “sleep position” as a quantitative/continuous variable (going from lower to higher levels of cuddling), then you’ll make a scatterplot. In either case, be sure your axis labels are crystal clear.  
2. Ask three questions to help you evaluate the statistical, construct, and external validity of this study. You don’t have to answer the question–just ask it. Use Chapter 8 to guide you on which questions to ask.  For example, for statistical validity you could ask, “how strong was the relationship between sleep position and stress? (i.e., what was the effect size?)”
3. Let’s discuss internal validity. Apparently sleep position (i.e., level of cuddling) is correlated with level of stress. Can we infer that cuddling reduces stress? Probably not from this study.The results show a correlation, so we have covariance. But the methods don’t support causation. For one thing, cuddling and stress were both measured around the same time, so there is no temporal precedence…we can’t tell which one came first in time.  What about internal validity?  We don’t have that either, because all of the variables were measured. Think of at least one specific third variable (“C”) that might be correlated with both high levels of cuddling (A) and low levels of stress (B). Explain how it might work. Here’s an example: Couples who are having more communication problems might have higher levels of stress (C –> B), and might also be less likely to cuddle (C –> A)  Now come up with your own “C” variable.

The original empirical article is available here (may be paywalled)