03. II unit

download 03. II unit

of 17

Transcript of 03. II unit

  • 8/13/2019 03. II unit

    1/17

    UNIT - IIEXPERIMENTAL DESIGN:

    EXPERIMENT RESEARCH DESIGN:

    An experiment is defined as manipulating (changing values/situations) one or moreindependent variables to see how the dependent variable(s) is/are affected, while alsocontrolling the affects of additional extraneous variables.

    Why do events occur some conditionsand not other conditions.

    Experimental research study involves intervention by the researcherbeyond reuirement for measurement. !sual intervention is to manipulate some variable in asetting and observe how it affects the sub"ect being studied.

    #esearcher manipulates the

    independent variable or explanatory variable and than observes whether hypothesi$eddependent variable is affected by the intervention.e.g. by standers and thieves % whether people observe (alone/company)

    LAB EXPERIMENT

    Experiments done in n rti!i"i# or "ontri$ed en$ironment%&'en "(se-nd-e!!e"t re#tions'ipbetween an independent and dependent variable of interestis to be "#er#) est*#is'ed, then all ot'er $ri*#esthat might contaminate or confound therelationship have to be ti+'t#) "ontro##ed. &t also is necessary to mnip(#te t'eindependent $ri*#e so t't t'e extent o! "s(# e!!e"ts "n *e est*#is'ed. Whencontrols and manipulations are introduced to establish cause'and'effect relationships in an

    artificial setting, is nown as lab experiments.Con!o(nded re#tions'ip When the dependent variable is not free from the influence ofextraneous variable, the relationship between the dependent and independent variables is saidto be confounded by an extraneous variables.

    Extrneo(s $ri*#e:&ndependent variables that are not related to the purpose of the study,but may affect the dependent variable are termed as extraneous variables. (E.g. children*sgains in social studies +- achievement and their self'concept +&-) % intelligence +Ex. -ar)

    C,NTR,L

    inimi$ing the effects of extraneous independent variables ' to restrainexperimental conditions % control the contaminating factors.e.g. 0# manager arrange for a special training to a set of newly recruited secretaries % to

    prove his boss % cause them to function more effectively % previous intermittent experience %learners previous experience has to be controlled.

    MANIPULATI,N

    &n order to examine the causal effects of an independent variable on adependent variable, certain manipulations need to be tried. &t means that create differentlevels of the independent variable to assess the impact on the dependent variable.

  • 8/13/2019 03. II unit

    2/17

    e.g. to test the effects of lighting on worer production levels among sewing machineoperators. 12'day period with the usual among of light % 34 watt lamps % 34 operators into 5groups 64 members each % manipulate the intensity of the light for the other two sub'groups %72'watt % 144'watt light'bulbs.' exposing different groups to different degrees of changes init.

    Tretment:8he different conditions under which experimental and control groups are putare usually referred to as 9treatments*. 8he manipulation of the independent variable is nownas the treatment, and results of the treatment are called treatment effects.

    e.g. the owner of toy shop % wages paid an hourly rate % whether paying piece rate increasetheir production levels. :trenuous ' tough % muscular strength % age, gender, prior experience(familiarity with the "ob) % ; groups of 12 people each % one to be used as control group.

    C,NTR,LLING THE C,NTAMINATING EX,GN,US ,R NUISANCE.

    /ARIABLES

    Mt"'in+atch the various groups by picing the confounding characteristics and

    deliberately spreading them across groups. E.g. if 64 women among 34 members, then eachgroup will be assigned 2 women. We are not sure that controlled all the nuisance factors,since we may not aware of them all. A safer bet is to randomi$e.

    Rndomi0tion

    Every member would have a nown and eual chance of being assigned to anyof these four groups (with no predetermination. 8hat is the variables of age, sex, and previousexperience % the controlled variables % will have an eual probability of being distributedamong the groups.

    increase in production)

    Experimental group 1 ? 1.44 per piece 14Experimental group 6 ? 1.24 per piece 12Experimental group 5 ? 6.44 per piece 64

  • 8/13/2019 03. II unit

    3/17

    in the causal effects (i.e. that variable causes variable B). Cield experiments have moreexternal validity but less internal validity (i.e. we cannot be certain of the extent to whichvariable alone causes variable B). Dote that in the lab experiment, the reverse is true. 8heinternal validity is very high but the external validity rather very low.

    8he lab setting does not reflect the real worldF setting, we do not now to what extent thelab findings validly represent the realities in the outside world.

    T1PES ,2 EXPERIMENTAL DESIGNS

    34In!orm# experiment# desi+ns:

    1. Gefore'and'after without control design

    8est area Hevel of phenomenon 8reatment Hevel of phenomenonbefore treatment () introduced after treatment (B)

    8reatment Effect I (B) ' ()6.After'only control design

    8est area 8reatment introduced Hevel of phenomenonafter treatment (B)

  • 8/13/2019 03. II unit

    4/17

    controlled artificial or lab setting, and the relationship has been established, they try to testthe causal relationship in a field experiment.

    2ACT,RS A22ECTING INTERNAL /ALIDIT1:ome confounding factors might still be

    present that could offer rival explanations as to what is causing the dependent variable. 8hesepossible confounding factors pose a threat to internal validity.

    Histor) E!!e"ts

  • 8/13/2019 03. II unit

    5/17

    &mproper or unmatched selection of sub"ects for the experimental and controlgroups. (E.g. impact of woring environment on employees* attitudes toward wor % somevolunteers might be lured through incentives)

    Sttisti"# Re+ression E!!e"ts

    8he members chosen for the experiment group have extremescores on the dependent variable.

    Mort#it) E!!e"ts

    Another confounding factor on the cause'and'effect relationship is themortality or attrition of the members in the experimental or control group or both, as theexperiment progresses.

    2ACT,RS A22ECTING EXTERNAL /ALIDIT1:ub"ect selection and its interaction with

    the treatment would also pose a threat to external validity.

  • 8/13/2019 03. II unit

    6/17

    e.g. effect of televised appearances on people*s votes for candidate % possible tomanipulate the timing or content or amount of televised political ads.

    @rganismic variable

    @rganismic (also called attribute) variables are ualities (i.e. religion,

    income, education, political party preferences, etc.) which persons already have. !nlieexperimental variables, these properties of sub"ects cannot be usually manipulated orexposed.

    Dependent $ri*#e 3D/4

    De!inition 8he variable that is primary interest to the researcher is called as dependentvariable.

    8he researcher has to understand and describe the dependent variable, or to explain itsvariability, or predict it. &t is possible to find a solution to the problem by finding whatvariables influence the dependent variable. 8he researcher is interested in uantifying andmeasuring the dependent variable and other variables that influence the dependent variable.

    Exmp#e

    1. &ntroduction of a product after test mareting % here, sales is the dependent variable.6. &nvestigating the debt euity ratio of a manufacturing company % ratio of debt to

    euity is the dependent variable.5. Employees are not loyal to the organi$ation % variance found in the level of

    organi$ational loyalty of employees. We need to find out what variable affect thevariance in loyalty.

    8here may be more than one dependent variable, for exmp#e

    1. 8here is always conflict between uality and volume of output.

    6. How cost production and customer satisfaction8he manager is interested to now the factors that influence all the dependent variables ofinterest and how some of them differ in regard to different dependent variables. 8heresearcher uses multivariate analysis in this regard.

    Independent $ri*#e 3I/4

    De!inition An independent variable is one that influences the dependent variable either apositive or negative way.

    Each unit independent variable increase, there is increase or decrease in the independentvariable, i.e. the variance in the dependent variable is accounted for by the independentvariable.

    Exmp#e !or independent $ri*#e

    iagram ' relationship between the independent variable (new product success) and thedependent variable (stoc maret price).

    Dew product :

    :uccess

    Independent $ri*#e Dependent $ri*#e

    :toc aret price

  • 8/13/2019 03. II unit

    7/17

    iagram of relationships between the independent variable (managerial values) and thedependent variable (power distance).

    anagerial values

    Independent $ri*#e Dependent $ri*#e

    Independent $ri*#e Dependent $ri*#e : S)non)ms

    Kredictor

  • 8/13/2019 03. II unit

    8/17

    Inter$enin+ $ri*#e 3I//4

    efinition 8he intervening variable surfaces between the times the independent variablesstart operating to influence the dependent variable and the time their impact is felt on it.8here is a temporal uality or time dimension to the intervening variable. 8he interveningvariable surfaces as a function of the independent variable(s) operating in any situation, andhelps to conceptuali$e and explain the influence of the independent variable(s) on thedependent variable.

    8he variable mentioned with regard to causal relationships are concretean clearly measureable then can be seen, counted or observed in some way. &-- may bedefined as that factor which theoretically affects, the observed phenomenon but cannot beseen, measured or manipulated

    e.g. A promotion campaign (&-) will increase savings activity (-), especially when freepri$es are offered (-), but chiefly among small savers (E-'control). 8his results fromenhancing the motivation to same (&--)

  • 8/13/2019 03. II unit

    9/17

    MEASUREMENT ,2 /ARIABLES : ,PERATI,NAL DE2INITI,N

    AND SCALES

    H,& /ARIABLES ARE MEASURED

    @b"ects that can be p')si"##) mes(red *) some "#i*rted instr(mentspose no

    measurement problem.e.g. length O width of rectangular table % with a measuring tape emographic characteristics of the office personnel % by simple uestions.

    ,PERATI,NAL DE2INITI,N:

    Dimensions nd E#ements o! Con"epts

    @perationali$ing, operationally defining a concept to render it measurable, isdone by looing at the behavioural dimensions, facets, or properties denoted by the concept.8hese are then translated into observable and measurable elements so as to develop an indexof measurement of the concept.

    e.g. @perationali$ing, the concept of Achievement otivation

    THE 2,UR T1PES ,2 SCALES

    SCALES

    A scale is a tool or mechanism by which individuals are distinguished as to how theydiffer from one another on the variables of interest to our study.

    8he higher the level of measurement, the more precision in measurement and everylevel up contains all the properties of the previous level. 8he four levels of measurement,from lowest to highest, are as follows

  • 8/13/2019 03. II unit

    10/17

    Nomin# S"#e

    A nominal scale is one that allows the researcher to assign sub"ects to certaincategories or groups.

    e.g. male and female (code 1 or 6) American, Australian,

  • 8/13/2019 03. II unit

    11/17

    traffic violations and the number of incorrect letters in a page of type script represent scoreson ratio scales.

    Eg. A person weighing 624 pounds is twice as heavy as one who weighs 162 pounds, notethat multiplying or dividing both of these numbers by any given number will preserve the

    ratio of 61.#atio scale represents the actual amounts of variables. easures of

    physical dimensions such as weight, height, distance, etc, are examples.

    8hus, proceeding from the nominal scale (the least precise type ofscale) to ratio scale (the most precise), relevant information is obtained increasingly.

    MEASUREMENT: SCALING< RELIABILIT1< /ALIDIT1

    SCALING TECHNI=UES 2RE=UENTL1 USED

    RATING SCALES

    Di"'otomo(s S"#e

    8he dichotomous scale is used to elicit a Bes or Do answer. Dote that anominal scale is used to elicit. 8he response.Eg. o you own a carM Bes / DoCte+or) S"#e

    8he category scale uses multiple items to elicit a single response. 8his alsouses the nominal scale.

    Eg. Where in

  • 8/13/2019 03. II unit

    12/17

    N(meri"# S"#e

    8he numerical scale also semantic differential scale, with the difference thatnumbers on a 2'point or 7'point scale provided, with bipolar ad"ectives at both ends. 8his isalso an interval scale.Extremely Extremely

    Kleased 7 3 2 ; 5 6 1 ispleased

    Itemi0ed Rtin+ S"#e

    A 2'point or 7'point scale with anchors, as needed, is provided for eachitem and the respondent states the appropriate number on the side of each item, or circles therelevant number against each item. 8his uses an interval scale.

    Eg. #espond to each item using the scale below, and indicate your response number on theline by each item.

    1 6 5 ; 2

    -ery !nliely !nliely Deither !nliely Hiely -ery Hiely Dor Hiely

    LLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLL

    1. & will be changing my "ob within the next 16 months LLLLLL6. & will tae on new assignments in the near future LLLLLL5. &t is possible that & will be out of this organi$ation within the next 16 months LLLLLL

    Dote that the above is balanced rating scale with a neutral point

    2ixed or Constnt S(m Rtin+ S"#e

    8he respondents are here ased to distribute a given number ofpoints across various items. 8his is more in the nature of an ordinal scale.Eg. &n choosing toilet soap, indicate the importance you attach to each of the following fiveaspects by allotting points for each to total 144 in all.

    Cragrance LLLL

  • 8/13/2019 03. II unit

    13/17

    :tate how would you rate your supervisor*s abilities with respect of each characteristics

    R 5 R 5 R 5 R 6 R 6 R 6

    R 1 R 1 R 1 Adopting odern Kroduct &nterpersonal 8echnology &nnovation :ills '1 '1 '1 '6 '6 '6 '5 '5 '5

    Grp'i" Rtin+ S"#e

    A graphical representation helps the respondents to indicate on this scale theiranswers to a particular uestion by placing a mar at the appropriate point on the line. 8his isan ordinal scale.

    14 Excellent@n a scale of 1 to 14how would you rate 2 All rightyour supervisorM

    1 -ery badConsens(s S"#e

    :cales are also developed by consensus (agreement), where a panel of "udgesselects certain items, which in its view measure the relevant concept. 8he items are chosen

    particularly based on their pertinence or relevance to the concept. @n such consensus scale isthe T'(rstone E>(# Apperin+ Inter$# S"#e, where a concept is measured by a complex

    process followed by a panel of "udges.

    ,t'er S"#es

    8here are also some advanced scaling methods such as multidimensionalscaling, where ob"ects, people, or both, are visually scaled, and a con"oint analysis is

    performed.

    #ADN&D= :

  • 8/13/2019 03. II unit

    14/17

    Gut when there are more than two stimuli to "udge, the number of "udgements reuired in apaired comparison is given by the formula

    D I n (n % 1) '''''''''''

    6

    #esponse Katterns of 144 embers* Kaired

  • 8/13/2019 03. II unit

    15/17

    ATTITUDE MEASUREMENT AND SCALES

    Among all the decisions, decisions taen in the mareting function are heavily based onattitude of the current / potential customers of a company.

    Attitude is a learned tendency of individuals to respond in a consistent manner to a givenob"ect of orientation

    E#ements o! mes(rin+ ttit(de re s !o##o?s

    Kersons, ob"ects, events, concepts or states to be observed

    Environmental conditions of the study

    :teps to be performed in collecting the observations

    &nstruments to be used to perform the steps

    @bservations to be collected

    2(ndment#s o! Dt Co##e"tion Met'ods

    /ri*i#it) met'ods

    Kaired comparison techniue#aning method@rdered'category sorting#ating methods

    =(ntitti$e @(d+ment met'ods

    irect "udgment methodsCraction method

  • 8/13/2019 03. II unit

    16/17

    G,,DNESS ,2 MEASURES /#idit) Re#i*i#it)&t is important to mae sure that the instrument that we develop to

    measure a particular concept is indeed accurately measuring the variable, and that in fact, weare actually measuring the concept that we set out to measure.

    8his ensures that in operationally defining

    perceptual and attitudinal variables, we have not overlooed some important dimensions andelements or included some irrelevant ones.

    8he use of better instruments will ensure moreaccuracy in results, which in turn, will enhance the scientific uality of the research. 0ence,in some way we need to assess the goodnessF of the measures developed.

    Item An#)sis

    &tem analysis is done to see if the items in the instrument belong there or not.Each item is examined for its ability to discriminate between those sub"ects whose totalscores are high, and those with low scores.

    RELIABILIT1

    8he reliability of a measure is an indication of the stability and consistencywith which the instrument measures the concept and helps to assess the goodnessF of themeasure.

    St*i#it) o! Mes(res8he ability of a measure to remain the same over time % despite

    uncontrollable testing conditions or the state of the respondents themselves'is indicative of itsstability and low vulnerability to changes in the situation

    Test-Retest Re#i*i#it)8he reliability coefficient obtained with a repetition of the same measure on a secondoccasion is called test'retest reliability

    E.g. same uestionnaire to same respondents after several wees or 3 months

    Pr##e#-2orm Re#i*i#it)

    8wo comparable sets of measures tapping the same construct are highly correlated. Gothforms have similar items and the same response format, the only changes being the wordingsand the order or seuence of the uestions.

    Intern# Consisten") o! Mes(res

    8he internal consistency of measures is indicative of the homogeneity of the items inthe measure that tap the construct. &n other words, the items should hang together as a setF,and be capable of independently measuring the same concept so that the respondents attachthe same overall meaning to each of the items.

    Sp#it-'#! re#i*i#it) "oe!!i"ient. 8his reflects the correlations between two halves of a set ofitems. 8he coefficients obtained will vary depending on how the scale is split. We obtainsplit'half reliability to test for consistency when more than one scale, dimensions, or factor, isassessed. We split the items across each of the dimensions or factors based on some

    predetermined logic.

    Inter-item Consisten") Re#i*i#it)#espondents* answer to all the items in a measure correlated with one another

  • 8/13/2019 03. II unit

    17/17

    /#idit)A characteristic of measurement concerned that a test measures what the researcher

    actually wishes to measureQ that differences found with a measurement tool reflect truedifferences among respondents drawn from a population. &t is concerned above the issue ofauthenticity of the cause'and'effect relationships (internal validity), and their generali$ability

    to the external environment (external validity).

    Content /#idit) 3Lo+i"# /#idit)4

    Adeuate and representative set of items that tab the concept. &t is a function of howwell the dimensions and elements of a concept delineated.

    2"e /#idit) is considered by some as a basic and very minimum index of content validity.Cace validity indicates that the items that are intended to measure a concept, do on the face ofit loo lie they measure the concept.

    Criterion-Re#ted /#idit)