Drriel: 1 revision imported

2024-02-09T00:10:18Z

1 revision imported

← Older revision	Revision as of 00:10, 9 February 2024
(No difference)

Drriel: Created page with "The '''identification of relationships''' is a family of research questions in evaluation and assessment that investigates whether two or more concepts, ideas, or variables are related to each other. Relationships can infer that when one thing is present, the other is also consistently present (or conversely not present) as well. This is often referred to as '''correlation''' in the evaluation field. == Definition == '''A relationship..."

2023-06-16T02:14:07Z

Created page with "The '''identification of relationships''' is a family of research questions in evaluation and assessment that investigates whether two or more concepts, ideas, or variables are related to each other. Relationships can infer that when one thing is present, the other is also consistently present (or conversely not present) as well. This is often referred to as '''correlation''' in the evaluation field. == Definition == '''A relationship..."

New page

The '''identification of relationships''' is a family of [[Research question|research questions]] in [[evaluation and assessment]] that investigates whether two or more concepts, ideas, or [[Data|variables]] are related to each other. Relationships can infer that when one thing is present, the other is also consistently present (or conversely not present) as well. This is often referred to as '''correlation''' in the evaluation field.

== Definition ==
'''A relationship''' between two concepts or variables in [[Evaluation and assessment|evaluation]] is the degree to which when one item is present, the other will also consistently have a related value. A demonstrated relationship between two concepts or variables is called '''correlation''' in the evaluation field.

A common correlation is a '''positive correlation''', in which the higher one item's value, the other item will consistently also have a higher value as well. However, other correlations exist. A '''negative correlation''' is when one item measures high, the (negatively) correlated item consistently measures low. Correlations can also indicate presence of items and not levels of measure, such as when one item is present, another will be consistently present (or not present, in the case of negative correlation).

== Additional Information ==
Evaluators should identify relationships between [[Data|variables]] (such as concepts, measures, and types of learners) within designs as to understand the dynamics between variables and to help identify potential causal mechanisms that lead to [[Learning objectives and outcomes|learning outcome achievement]] and the [[Participation|use of educational products]]. If one variable, when it is measured high for a learner shows another variable being consistently high as well, it could indicate important aspects of how an educational product works.

In the case of correlation, variables can represent any concept, idea, or behavior, such as a person's level of [[Behavior of the learner (KAB)|behavior]], their [[Participation|level of participation]] in the learning environment, their measured [[Competency vs. learning|achievement or competence]] with the knowledge and skills of the learning, their [[Attitudes and affect of the learner (KAB)|affect or perspectives]], or any number of environmental or technology factors. The important idea with correlation research is to ''identify'' the concept being studied, figure out a way how to ''measure it or categorize it'' from the learning environment, and ''conduct an analysis on it.''

Three general types of correlation studies can be conducted to see what kinds of relationships exist between two variables:

* '''Positive correlation.''' This type of study investigates whether there is a link between two variables, with both consistently being present and scaling to higher levels when one or the other is observed to be higher. In this case, when one is higher, the other is also higher.
* '''Negative correlation.''' This type of study tests if the evaluators observe the presence or high levels of one variable, they will see a consistent absence or low levels of another variable. When one is higher, the other is lower or nonexistent.
* '''No correlation'''. This study identifies that there is no correlation between two concepts or variables. This is useful if there is a hypothesized or assumed relationship between two concepts, but is revealed there is actually no observed relationship.

It is important to note that scientific analysis practice emphasizes that '''correlation does not imply causality'''. This simply means that because two concepts or measures are ''related'' does not necessarily mean that one causes the other to be that way. The relationship could be coincidential, or the relationship could be caused by a third concept that is not considered by the evaluator. In any regard, a relationship should be interpreted as just that - a relationship. A statitstically significant relationship between two variables only states that the two variables consistently "run together" or have values that consistently operate in the same direction - '''not''' that one variable ''causes'' the other to be higher or lower. Studying correlation and relationships is valuable to an evaluator, however, as it can reveal some of the causal mechanisms that might be in place between the two variables. Additional more rigorous studies can be conducted to identify and determine '''causality''' between the variables if this is desired.

===== Common methods for identifying relationships =====

* '''[[Quantitative analysis|Quantitative]] correlation studies''' that examine [[Data|datasets]] for consistency between two measured ''[[Data|continuous (scaled)]]'' variables, such as when one is high the other is high (positively correlated), or when one is high the other is low (negative correlation). The most common statistical method used is '''Pearson's r.''' More complex statistical relationship identification methods can also be used for specific contexts to address limitations or unique structures of datasets to recieve more reliable results.
* '''[[Qualitative analysis|Qualitative studies]]''' that [[Coding (analysis)|code]] for the presence and degree of concepts, behaviors, or other items can be systematically investigated to see how, why, and with what mechanisms two or more concepts are consistently ''together'' in a learning context. Researchers can investigate participants' perceptions, their dialogue, and other [[Interactions and activities|interactions]] to see the degree to which concepts "run together" and are equally present in activities (or, conversely, not present in activities). A range of qualitative methods can be used to investigate relationships between concepts, including '''basic qualitative analysis, process study, case study, grounded theory,'''

== Tips and Tricks ==

* Consider what variables you may test for whether there is an observed relationship or correlation between them. By finding that one variable's high measure also indicates that another variable will be high can be helpful for evaluators to identify things they should focus on and can reveal parts of the learning experience that are particularly valuable to achievement. For instance, if a relationship is identified between discussion participation and final grade, where higher levels of participation in the discussion are correlated or linked with a higher final grade in the course, such a relationship may reveal an important contribution that the discussion is having in the overall learning experience. It doesn't not necessarily mean that discussions lead to (or cause) better learning outcomes. Instead, it reveals that something important is happening with discussions and it merits further evaluation to see what may be causing higher levels of achievement.
* Remember that correlation or a relationship between two variables or concepts does not necessarily mean that one causes the other to be that way. This is an easy trap to fall into, as when people hear that when one thing is higher, so is the other, and thus jump to the conclusion that the first item's high measured value is causing the second item to also be high. However, correlation simply measures for the consistency with which both items are present in the case of positive correlation, or not present in the case of a negative correlation.

== Related Concepts ==

* [[Evaluation and assessment]]
* [[Research question]]
* [[Analysis method]]
* [[Data]]

== Examples ==

None yet - check back soon!

== External Resources ==

None yet - check back soon!

Identification of relationships (analysis) - Revision history

Drriel: 1 revision imported