up:: [[HD 2830 Research Methods and Design]] Tags:: #🌱 # Interrogating null effects Getting null effects in a study can suck. Most researchers think it's more exciting to show something changes because of something than to show there is no difference. But these effects are still essential to report. And you should report them as not doing so would violate the [[Disinterestedness Mertonian Norm]]. When a study shows that their is a null effect it doesn't necessarily mean that the treatment doesn't have an effect. It could be a result of bad study design. Here are three situations in which bad study design could lead to a null effect: ### Weak manipulations Weak manipulations are when the study designer doesn't make the independent variable strong enough. As a result there is no effect when there would be at a higher degree of manipulation. ### Insensitive measures Insensitive measures occur when the dependent variable isn't sensitive enough to changes. This could result in no effect being seen when a more sensitive instrument might have found something. ### Ceiling and floor effects Ceiling effects occur when all of the scores are squeezed together at the high end where as floor effects occur when all scores are squeezed together at the low end. ### Manipulation Checks are the Solution Manipulation checks are separate dependent variables experimenters include in studies to make sure the manipulation worked. They fight against ceiling and floor effects as well as insensitive measures. For example, in the anxiety study, after telling people they were going to receive a 10-volt, 50-volt, or 100-volt shock, the researchers might have asked: How anxious are you right now, on a scale of 1 to 10? ## Other reasons for Null Effect ### Design confounds A aspect of the study design influences scores where there shouldn't be. For example, an experimenter that is grouchy when everyone else isn't. ### High Within Groups Variability High within groups variability distracts from the ability to show between groups variability. One way to take this into account is to use a [[Types of Experimental Design#Within Subject Design|within subject design]] to get rid of individual differences. ### Situation Noise The situation a study is done in might cause systematic differences that shouldn't be present in some scores. Solution: Carefully control study environment. Related: ___ # Resources