Very common in experimental research, treatment and control groups allow researchers to compare and contrast the impact of a specific condition, drug/compound, intervention, etc. aka “treatment” in question. In establishing control groups, all factors should be as equal as possible between the two groups with only one exception, the treatment (condition, drug/compound, interventions, etc. under study).

Very common in quantitative research, positive and negative controls allow the researcher to be certain of two critical points.

One, that the experiment, as designed and carried out, is capable of detecting/demonstrating the result/condition/reaction, etc. under question. This comes from using a positive control; adding a factor known to produce a positive result with all other conditions being the same. When a positive control works, it allows researchers to trust that any negative results for the experiment itself are not false negatives.

Two, that the experiment, as designed and carried out, is not contaminated leading to false positives. This comes from using a negative control; running the experiment without a sample or known positive control so that if any positive results are detected/demonstrated they can only be the result of contamination. Thus, when negative controls work, researchers can trust that any positive results for the experiment itself are not false positives.

It is important to note that for any experimental design for which positive/negative controls are needed, every run of the experiment will need positive and negative controls built in otherwise any results without them are questionable.

Sometimes in research the best experimental designs are still only indirect measures of what researchers are testing. In this situation, it is especially important to utilize multiple, independent* methods to test for the same information. When more than one independent method indicates the same result or conclusion, researchers can be more certain of the outcomes.

*Independent methods are ones that are completely separate from one another. For example, if researchers wanted to test ice cream preferences in a population, two independent methods would be 1) using a survey to collect personal accounts and 2) analyzing the purchase records from stores that sell ice cream.

Very common in qualitative research, but also a critical part of any background literature review is examining multiple sources of information to ensure the full wealth of knowledge has been examined. When utilizing multiple sources for information, it’s important to cover a wide breadth as well as a sufficient depth.

When talking with people or observing populations (past or present), researchers should seek information from a wide range of individuals or sources. Using multiple sources (e.g. interviews, historical records, etc.) allows qualitative researchers to triangulate or cross-verify their findings.

When conducting a literature review, researchers should use a wide range of keywords to find articles applicable to your study. Researchers should also make sure to honestly consider articles discussing results that conflict with their own results or assumptions.

Two common controls in qualitative research, fact checking and member checking allow researchers to introduce rigor into the analysis of people’s stories. Even when opinions are shared, researchers will still have a set of facts that can be verified to confirm the trustworthiness of the source and the accuracy of the information capturing process.

For example, researchers can fact check comments on such things as: the weather; what other witnesses or participants reported; dates and times of known or cataloged occurrences; recordings from the event, day, or time period; and even other documented comments shared by the subject.

Member checking comes into play when researchers reach back out to participants to confirm the accuracy of their notes, transcripts, and their interpretation, as well as verifying the details that can and cannot be shared in places such as publications or reports.

Perhaps the most critical control of all, self checking is where researchers honestly consider their intentions, motivations, and preconceived notions plus the impacts each of these can have on the research in question. Qualitative researchers often make note of these during the data collection and analysis processes (commonly called reflexivity or positionality), however, researchers engaged in any type of research should make use of this control.

Anytime there is a possibility that a subject’s and/or data collector’s knowledge of the experiment or treatment conditions can impact the results, blind or double-blind controls should be used. These allow researchers to alleviate those concerns.

In blind controls, the subjects are deliberately made unaware as to which group they’ve been assigned and/or what the researchers are measuring. This way, the subject’s assumptions or expectations about either the treatment or control scenario are not impacting their behavior or responses.

In double-blind controls, the same blind control set up exists with the added condition that the data collector is also unaware as to which group the subject has been assigned. This way the data collectors cannot consciously or unconsciously influence either the subjects themselves or the data interpretation/analysis process.

In both quantitative and qualitative research a population under study can be too large to test or examine each individual. Researchers in this situation have to rely on random sampling to cover the range of natural differences within the given population, and a large enough sampling size to ensure adequate representation of the same population.

Researchers should factor statistics (power analysis), logistics, and resources when determining an appropriate sample size. Special steps should also be taken in the methods to make sure the sampling is random. See some sample questions below for consideration:

• With human subjects for example, where is the recruitment information posted?
  • What language is used to draw in participants?
• With plant and animal subjects, from where were they sampled?
  • What methods were used to capture or collect the samples?
• Is it possible that there are variations within the population that would preclude a subset of individuals from being picked up by the sampling methods?
  • What steps could correct that imbalance?