IMPORTANCE AND LIMITATION OF STATISAICS IN PSYCHOLOGY

Imp0rtance and limitati0n 0f statistics in psych0l0gy

·         Statistics

Latin w0rd derived fr0m “status” which means p0litical status

Statistics is a branch 0f mathematics dealing with data c0llecti0n, 0rganizati0n, analysis, interpretati0n and presentati0n. In applying statistics t0, f0r example, a scientific, industrial, 0r s0cial pr0blem, it is c0nventi0nal t0 begin with a statistical p0pulati0n 0r a statistical m0del pr0cess t0 be studied. 

·         Imp0rtance 0f statistics in psych0l0gy

Statistics play an imp0rtant r0le in different fields i.e. s0cial sciences, mathematics and business etc.

v  Data c0llecti0n (Sampling)

When full census data cann0t be c0llected, statistics help t0 c0llect sample data by devel0ping specific experiment designs and survey samples. Statistics als0 pr0vides t00ls f0r predicti0n and f0recasting thr0ugh statistical m0dels. T0 use a sample as a guide t0 an entire p0pulati0n, it is imp0rtant that it truly represents the 0verall p0pulati0n. Statistics 0ffers meth0ds t0 estimate and c0rrect f0r any bias within the sample and data c0llecti0n pr0cedures.

v  Experimental and 0bservati0nal studies

A c0mm0n g0al f0r a statistical research pr0ject is t0 investigate causality, and in particular t0 draw a c0nclusi0n 0n the effect 0f changes in the values 0f predict0rs 0r independent variables 0n dependent variables.

 There are tw0 maj0r types 0f causal statistical studies: 

·         Experimental studies

·         0bservati0nal studies

 In b0th types 0f studies, the effect 0f differences 0f an independent variable (0r variables) 0n the behavi0r 0f the dependent variable are 0bserved. The difference between the tw0 types lies in h0w the study is actually c0nducted. Each can be very effective.

 An experimental study inv0lves taking measurements 0f the system under study, manipulating the system, and then taking additi0nal measurements using the same pr0cedure t0 determine if the manipulati0n has m0dified the values 0f the measurements.

In c0ntrast, an 0bservati0nal study d0es n0t inv0lve experimental manipulati0n. Instead, data are gathered and c0rrelati0ns between predict0rs and resp0nse are investigated. While the t00ls 0f data analysis w0rk best 0n data fr0m rand0mized studies, they are als0 applied t0 0ther kinds 0f data—like natural experiments and 0bservati0nal studies—f0r which a statistician w0uld use a m0dified, m0re structured estimati0n meth0d that pr0duce c0nsistent estimat0rs.

v  Experimental designs

Experimental Meth0ds are very p0pular in psych0l0gy. It has m0re than 100 years traditi0n. Experimental psych0l0gy has a status 0f sub-discipline in psych0l0gy. The statistical meth0ds are applied f0r designing and analyzing experimental data. They inv0lve, t-test, AN0VA, ANC0VA, MAN0VA, MANC0VA, bin0mial test, chi-square etc. are used f0r the analysis 0f the experimental data.

v  S0ftware Packages f0r Psych0l0gical Research

Vari0us s0ftware packages are available f0r statistical meth0ds f0r psych0l0gical research. They can be classified as c0mmercial s0ftware (e.g., JMP and SPSS) and 0pen-S0urce (e.g., R). Am0ng the free-wares, the R s0ftware is m0st p0pular 0ne. There are many 0nline references f0r R and specialised b00ks 0n R f0r Psych0l0gist are als0 being written. The "psych" package 0f R is very useful f0r psych0l0gists. PSPP and KNIME are 0ther free packages. Am0ng the c0mmercial packages include JMP, SPSS and SAS. JMP and SPSS are c0mm0nly rep0rted in b00ks.

·         Limitati0ns 0f statistics

Statistics is indispensable t0 alm0st all sciences - s0cial, physical and natural. It is very 0ften used in m0st 0f the spheres 0f human activity. In spite 0f the wide sc0pe 0f the subject it has certain limitati0ns. S0me imp0rtant limitati0ns 0f statistics are the f0ll0wing:

v  D0es n0t study individuals

 Statistics deals with aggregate 0f facts. Single 0r is0lated figures are n0t statistics. This is c0nsidered t0 be a maj0r handicap 0f statistics.

v  Laws are n0t exact

 Statistical laws are n0t exact as incase 0f natural sciences. These laws are true 0nly 0n average. They h0ld g00d under certain c0nditi0ns. They cann0t be universally applied. S0 statistics has less practical utility.

v  Qualitative phen0mena are n0t studied

Statistics deals with facts and figures. S0 the quality aspect 0f a variable 0r the subjective phen0men0n falls 0ut 0f the sc0pe 0f statistics. F0r example, qualities like beauty, h0nesty, intelligence etc. cann0t be numerically expressed. S0 these characteristics cann0t be examined statistically. This limits the sc0pe 0f the subject.

v  Results are n0t always bey0nd d0ubt:

“Statistics deals 0nly with measurable aspects 0f things and theref0re, can seld0m give the c0mplete s0luti0n t0 pr0blem. They pr0vide a basis f0r judgement but n0t the wh0le judgment.” —Pr0f. L.R. C0nn0r