Scribd Logo
Upload

Welcome to Scribd!

  • 39 views

    Factor Analysis

    Uploaded by

    Mary Rose Linihan
    Factor analysis is a statistical method used to describe variability among observed variables in terms of a potentially lower number of unobserved variables called factors. It can reduce the dimensionality of the data while retaining as much information as possible. The method works by finding clusters of highly correlated observed variables and representing them with factors. Researchers use factor analysis to investigate concepts that cannot be directly measured, like socioeconomic status or leadership style, by collapsing many variables into a few interpretable underlying factors. Sample sizes need to be large, usually in the hundreds, for stable results, though fewer cases may suffice if the data are well-behaved. Factor analysis requires subjective judgments and interpretations of results can vary.

    Copyright:

    © All Rights Reserved
    Download as DOC, PDF, TXT or read online from Scribd

    Factor Analysis

    Uploaded by

    Mary Rose Linihan
    39 views4 pages
    Factor analysis is a statistical method used to describe variability among observed variables in terms of a potentially lower number of unobserved variables called factors. It can reduce the dimensionality of the data while retaining as much information as possible. The method works by finding clusters of highly correlated observed variables and representing them with factors. Researchers use factor analysis to investigate concepts that cannot be directly measured, like socioeconomic status or leadership style, by collapsing many variables into a few interpretable underlying factors. Sample sizes need to be large, usually in the hundreds, for stable results, though fewer cases may suffice if the data are well-behaved. Factor analysis requires subjective judgments and interpretations of results can vary.

    Copyright

    © © All Rights Reserved

    Available Formats

    DOC, PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    Factor analysis is a statistical method used to describe variability among observed variables in terms of a potentially lower number of unobserved variables called factors. It can reduce the dimensionality of the data while retaining as much information as possible. The method works by finding clusters of highly correlated observed variables and representing them with factors. Researchers use factor analysis to investigate concepts that cannot be directly measured, like socioeconomic status or leadership style, by collapsing many variables into a few interpretable underlying factors. Sample sizes need to be large, usually in the hundreds, for stable results, though fewer cases may suffice if the data are well-behaved. Factor analysis requires subjective judgments and interpretations of results can vary.

    Copyright:

    © All Rights Reserved
    Download as DOC, PDF, TXT or read online from Scribd
    Download as doc, pdf, or txt
    39 views4 pages

    Factor Analysis

    Uploaded by

    Mary Rose Linihan
    Factor analysis is a statistical method used to describe variability among observed variables in terms of a potentially lower number of unobserved variables called factors. It can reduce the dimensionality of the data while retaining as much information as possible. The method works by finding clusters of highly correlated observed variables and representing them with factors. Researchers use factor analysis to investigate concepts that cannot be directly measured, like socioeconomic status or leadership style, by collapsing many variables into a few interpretable underlying factors. Sample sizes need to be large, usually in the hundreds, for stable results, though fewer cases may suffice if the data are well-behaved. Factor analysis requires subjective judgments and interpretations of results can vary.

    Copyright:

    © All Rights Reserved
    Download as DOC, PDF, TXT or read online from Scribd
    Download as doc, pdf, or txt
    of 4

    TOPICS

    Introduction to factors analysis


    - basic concepts and common examples
    -Factor analysis according to Adelman and Morris (1991), is a way to condense the data
    in many variables into a just a few variables. Sometimes called "dimension reduction".
    You can reduce the dimension of your data into one or more super-variables. This can be
    done by running PCA and keeping the first principal component (also known as factor).

    HOW FA CAN HELP RESEARCHERS


    - It is useful in
    1. Condensing variables
    2. Uncovering clusters of responses
    3. Providing simplicity after reducing variables

    THE ADVANTAGE OF PCA over an average is that it automatically weights each of the
    variables in the calculation.

    FACTOR ANALYSIS
    Acc. To Hatcher, 1991
    - is a method for modeling observed variables, and their covariance structure, in terms of
    a smaller number of underlying unobservable (latent) "factors"
    -it is generally an exploratory/descriptive method that requires many subjective
    Judgments.
    - It is widely used tool and ofetn controversial because the models, methods and
    subjectivity are so flexible that debates about interpretations can occur.
    - Done by means of principal components analysis (CPA)
    WHY USE FACTOR ANALYSIS
    - Acc to Hackett, 1986
    - is a useful tool for investigating variables relationships for complex concepts such as
    socioeconomic status, dietary patterns, or psychological scales
    - Allows researchers to investigate concepts that are not easily measured directly by
    collapsing a large number of variables into a few interpretable underlying factors.

    WHAT IS FACTOR
    -Key concept of factor analysis is that multiple observed variables have similar patterns
    of responses because they are all associated with a latent (not directly measured)
    variables.
    Ex. People may respond similarly to questions about income, education, and occupation,
    which are all associated with the latent variable socioeconomic status.

    FACTOR LOADINGS
    -Acc to Rahn 2001, it is a relationship of each variable to the underlying factor is
    expressed by the so- called factor loading. It can be interpreted like standardized
    regression coefficients (which denotes a correlation between two variables).
    -Once you run a factor analysis and think you have some usable results, its time to
    eliminate variables that are not strong enough. They are usually the ones w/ low factor
    loading.
    -As a rule of thumb, your variable should have a rotated factor loading of at least /0.4/
    onto one of the factors in oder to be considered important.
    -Some researchers use mush more stringent criteria such as cut off of absolute value of
    0.7

    FACTOR IN FACTOR ANALYSIS


    -Acc to Martin & Osborne, 2000. Factor analysis, a factor is a latent (unmeasured)
    variable that express itself through its relationship w/ other measured variables.
    Ex. A variable like leadership. We may want to measureba persons or an organizations
    leadership style, but this is the kind of construct that would be impossible to measure
    using a single variable. Its just too abstract and multifaceted, although it does represent a
    single concept.
    - The purpose of factor analysis is to analyze these patterns of response as a way of
    getting at this underlying factor.

    FACTOR SCORES
    -allow you to use a single variable as a measure of the factor in the other analyses, raher
    than a set of items.

    HOW BIG THE SAMPLE SIZE DO YOU NEED FOR FACTOR ANALYSIS
    -Acc to Martin & Harrington, 1999. Most of the time we plan the sample size for a data
    set based on obtaining reasonable statistical power for a key analysis of that data set. But
    nit only issue in sample size and not every statistical analysis uses p-value. One example
    is Factor Analysis.
    -Factor analysis is a measurement model of an underlying construct.
    - The focus is understanding w/c variables are associated with which latent constructs.
    The approach is slightly different if your running an exploratory or a confirmatory model,
    but the overall focus is the same.

    RULES OF THUMB

    TAKEAWAYS
    1. You are going to need a large sample. That means in the hndreds of cases. More is
    better.
    2. You can get away w/ fewer observations if the data are well-behaved. If there are no
    missing data and each variable highly loads on a single factor and not others, you wont
    need as many cases.
    3. The main issue w/ small data sets is overfitting (a secondary issue is if the sample is
    really small, the model wont even converge).

    - Yes, Confirmatory Factor Analysis can use p-values, for overall model fit chi-square test
    as well as specific path coefficients. Exploratory Factor Analysis does not.
    HOW MANY VARIABLES AND OBSERVATIONS

    Acc to Rahn, 2001.


    - Generally each factor should have at least three variables with high loadings.
    - It is important to have a sufficient number of observations to support your factor
    analysis: per variable should have about 20 observations in the data set to ensure stable
    results.

    ROTATIONS
    - Acc to Rahn 2001, axes of the factors can be rotated within the multidimensional
    variable space.
    -Ex. The program looks first for the strongest correlations between variables and the
    latent factor, and makes that factor 1 or called as axis 1. Then look for the second set of
    correlations calls it factor 2 or axis 2. And so on.
    - Inorder to make the location of the axes fit the actual data points better, the program can
    rotate the axes, Ideally, the rotation will make the factors more easily interpretable.
    - Rotations that allow for correlation are called obliques rotations; rotations that assume
    the factors are not correlated are called orthagonal rotations.
    -Programs offer many different types of rotations. An important difference between them
    is that they can create factors that are correlated or uncorrelated w/ each other.
    - If orthogonal ang rotation, we specified that factor 1 and 2 are not correlated.

    WHY DO VARIABLES NEED TO BE SELECTED BEFORE ANALYZING THE


    DATA.
    - It is important to select the variables before analyzing the data in order to determine
    what statistical test to be used such as multiple regression that easily come up with the
    results and avoid the biased judgment that can possibly happen. Factor analysis will
    nearly always produce a 'solution'. However, it may well be a nonsense solution. Factor
    analysis cannot distinguish between variables that are casually related and those that are
    non-casually related.

    HOW TO NARROW DOWN THE CHOICE OF VARIABLES


    -It is important to select the dependent and independent variables that can be used in
    research. State the dependent variable and the independent variable and likely the
    mechanism on why they are related. In factor analysis, think about the variables involve
    and the idea of what they might have in common. Avoid including variables where any
    correlation is more likely due to casual relationships than to the variables having
    something in common.

    EXPLORARTORY FACTOR ANALYSIS(EFA)


    -refers to a class of procedures that include centroid, principal components, and principal
    axis factor analysis, among many others, that differs in criteria used to derived factors
    (Kline, 2005). It does not require prior hypothesis
    -The general purpose is to identify the common factor and explain their relationship to
    the observed data.
    Validity- measure the thruthfullness and appropriateness of data. It is important especially
    in selecting an instrument that we need to considered the instrument being used was
    validated.
    Evidence based on text content- It is to analyze the relationship between the instruments
    content and the construct it is intended to measure.

    Evidence based on response process- Information on how respondents respond the


    instrument items

    Evidence based on internal structure - The internal relationships between instrument


    items and instruments components.

    Evidence based on relations to other variables- analyses of the relationships of instrument


    score to variables external to other instruments.

    Evidence based on the consequences of testing - The extent to w/c the consequences of
    the use of the score are congruent w/ the proposed uses of the instrument.

    You might also like