Descriptive Statistics, Part 2

The Doctoral Journey
26 Aug 201315:20

Summary

TLDREste tutorial explica cómo calcular estadísticas descriptivas en SPSS, centrándose en medidas de posición como percentiles y cuartiles. Se detalla cómo interpretar estos valores para entender la posición de una puntuación dentro de una distribución. Además, se introduce el concepto de puntuaciones estándar (z-scores), que miden la posición relativa a la media y su desviación estándar. Finalmente, se ofrecen instrucciones paso a paso para calcular estas estadísticas en SPSS utilizando funciones como Frequencies, Descriptives y Explore.

Takeaways

  • 📊 Los percentiles son valores que dividen un conjunto de observaciones en 100 partes iguales, indicando la porción de datos que se encuentra por debajo de un punto dado.
  • 📈 El percentil 50 representa el punto medio de una distribución, también conocido como mediana.
  • 🔢 Los cuartiles son valores que dividen un conjunto de datos en cuatro partes iguales, representados por Q1, Q2 (mediana) y Q3.
  • 📉 El Q1 se corresponde con el 25% percentile, el Q2 con el 50% percentile (mediana) y el Q3 con el 75% percentile.
  • 🎓 Los scores en el percentil 75 significan que el 75% de los individuos tienen puntajes por debajo de ese valor, y el 25% por encima.
  • 📚 Los scores estánndarizados, como el z-score, miden cuántas desviaciones estándar un puntaje está del promedio.
  • ⚖️ Un z-score de 0 indica que el puntaje está al mismo nivel que el promedio, mientras que valores positivos y negativos indican desviaciones por encima o por debajo del promedio, respectivamente.
  • 📉 Un z-score de 1 significa que el puntaje está una desviación estándar por encima del promedio, y así sucesivamente.
  • 🚫 Por convención, los puntajes con z-scores menores de -1.96 o mayores de 1.96 se consideran inusuales o extremos en una distribución normal.
  • 💡 Los z-scores pueden ayudar a comprender la posición relativa de un puntaje dentro de una distribución, lo cual puede ser útil para educadores y analistas de datos.

Q & A

  • ¿Qué son los percentiles y cómo se definen?

    -Los percentiles son valores que dividen un conjunto de observaciones en 100 partes iguales o puntos en una distribución por debajo de los cuales un cierto porcentaje de los casos yacen.

  • Si un individuo obtiene un puntaje del 33%, ¿qué significa esto en términos de percentiles?

    -Un puntaje del 33% en el 50º percentile indica que el 50% de las personas obtuvieron un puntaje menor que 33.

  • ¿Cómo se interpreta el 75º percentile en el contexto de un puntaje del 73%?

    -El 75º percentile significa que el 75% de las puntuaciones son menores que 73, y el 25% son mayores.

  • ¿Qué son los cuartiles y cómo se relacionan con los percentiles?

    -Los cuartiles son valores que ordenan los datos en cuatro partes iguales. El primer cuartil es igual al 25º percentile, el segundo (mediana) al 50º, y el tercero al 75º percentile.

  • Si un puntaje es igual al segundo cuartil, ¿qué significa eso?

    -Si un puntaje es igual al segundo cuartil, significa que está en el 50º percentile o mediana, y que el 50% de los puntajes están por debajo de ese puntaje.

  • ¿Qué es un puntaje de desviación estándar y cómo se calcula?

    -Un puntaje de desviación estándar, o z-score, indica cuántas desviaciones estándar un puntaje está del promedio. Se calcula restando el promedio y dividiendo por la desviación estándar.

  • ¿Cómo se interpreta un z-score de +2 en el contexto de una distribución de puntajes?

    -Un z-score de +2 indica que el puntaje está a dos desviaciones estándar por encima del promedio.

  • ¿Qué convención se sigue para considerar un z-score como inusual o extremo?

    -Por convención, los z-scores menores de -1.96 o mayores de 1.96 se consideran inusuales o extremos, ya que representan los 5% más extremos de la distribución.

  • ¿Cómo se pueden calcular los z-scores en SPSS?

    -En SPSS, para calcular z-scores se utiliza la función 'Descriptivos'. Se selecciona la variable a analizar y se marca la opción 'Guardar valores estandarizados como variables'.

  • ¿Cuáles son las diferencias entre las funciones 'Frecuencias', 'Descriptivos' y 'Explorar' en SPSS para calcular estadísticas descriptivas?

    -Las funciones 'Frecuencias', 'Descriptivos' y 'Explorar' en SPSS calculan estadísticas descriptivas como el promedio y la desviación estándar, pero difieren en que 'Frecuencias' permite calcular percentiles específicos y analizar una variable a la vez, mientras que 'Explorar' permite desagregar una variable por otra.

Outlines

00:00

📊 Descripción de Estadísticas Descriptivas en SPSS

Amanda rockinson-szapkiw nos presenta cómo calcular estadísticas descriptivas en SPSS. Explica inicialmente qué son los percentiles, que son valores que dividen un conjunto de observaciones en 100 partes iguales, y cómo se pueden usar para determinar la posición relativa de una observación dentro de una distribución. Por ejemplo, un puntaje de 33 podría estar en el percentil 50, lo que significa que el 50% de las observaciones están por debajo de este valor. También habla sobre cuartiles, que son valores que dividen un conjunto en cuatro partes iguales, y cómo estos están relacionados con los percentiles (por ejemplo, el cuartil 1 está en el percentil 25, el cuartil 2 en el percentil 50 y el cuartil 3 en el percentil 75).

05:02

📈 Medidas de Tendencia Central y Dispersión

En este párrafo, Amanda profundiza en la diferencia entre medidas de tendencia central y dispersión. Mientras que las medidas de tendencia central y dispersión ayudan a entender la distribución general, los percentiles y cuartiles ayudan a entender la posición específica de una puntuación dentro de esa distribución. Luego, introduce las puntuaciones estándar o 'z-scores', que miden cuántas desviaciones estándar se encuentra un puntaje con respecto a la media. Se explica cómo calcular un z-score y cómo interpretar sus valores, ya sea que sean positivos o negativos, y cómo estos valores pueden indicar si un puntaje es considerado usual o extremo en una distribución normal.

10:02

👤 Ejemplos Prácticos de Z-Scores

Amanda utiliza ejemplos prácticos para ilustrar cómo se pueden interpretar los z-scores. Primero, menciona una distribución de puntajes de exámenes de estadísticas educativas con una media de 100 y una desviación estándar de 15. Calcula el z-score de un estudiante que obtuvo un 130, que resulta ser de +2, lo que significa que el puntaje está dos desviaciones estándar por encima de la media. Luego, aplica la misma lógica al ejemplo de la altura de las mujeres en los Estados Unidos, que tiene una media de 64 pulgadas y una desviación estándar de 2.56 pulgadas. Calcula los z-scores para mujeres de 66 y 71 pulgadas, respectivamente, y concluye que la altura de 71 pulgadas podría considerarse extrema.

15:04

💻 Cómo Calcular Z-Scores en SPSS

Finalmente, Amanda explica cómo calcular z-scores y otras estadísticas descriptivas en SPSS utilizando tres funciones principales: frecuencias, descriptivas y explorar. Detalla que la función descriptivas es la que permite calcular z-scores al marcar la opción de 'guardar valores estandarizados como variables'. Asegura que, al final de este tutorial, el usuario debería poder identificar y proporcionar ejemplos de diferentes medidas de posición, como percentiles, cuartiles y z-scores, y entender cómo calcular estas estadísticas en SPSS.

Mindmap

Keywords

💡Percentiles

Los percentiles son valores que dividen un conjunto de observaciones en 100 partes iguales. En el vídeo, se usa el ejemplo de una distribución de puntajes donde un puntaje de 33 es igual al percentil 50, lo que significa que el 50% de las personas obtuvieron puntajes por debajo de 33. Esto ayuda a entender la posición relativa de un puntaje dentro de una distribución.

💡Quartiles

Los quartiles son valores que dividen un conjunto de datos en cuatro partes iguales. Se mencionan en el vídeo como una forma de entender la distribución de datos, donde el primer quartil es igual al 25%, el segundo al 50% (mediana) y el tercer al 75%. Por ejemplo, un puntaje de 33 que es igual al 50%, también es igual al segundo quartil.

💡Mediana

La mediana es el punto central de una distribución de datos, también conocido como el segundo quartil. Se menciona en el vídeo que el puntaje de 33 es igual a la mediana, lo que indica que es el punto de referencia para la mitad superior e inferior de la distribución de puntajes.

💡Z-score

El Z-score es una medida estándar que indica cuántas desviaciones estándar un punto de datos está por encima o por debajo de la media. En el vídeo, se explica que un Z-score de +2 significa que el puntaje está dos desviaciones estándar por encima de la media, como en el caso de un estudiante que obtiene 130 puntos en una prueba con una media de 100 y una desviación estándar de 15.

💡Desviación estándar

La desviación estándar es una medida que indica la dispersión de los datos alrededor de la media. Se menciona en el contexto de calcular el Z-score, donde se divide el puntaje por la desviación estándar para determinar cuántas desviaciones estándar está el puntaje por encima de la media.

💡SPSS

SPSS es un software estadístico mencionado en el vídeo para calcular estadísticas descriptivas, como percentiles, quartiles y Z-scores. Se describe cómo usar diferentes funciones de SPSS para analizar datos y entender la distribución y la posición de los puntajes dentro de una muestra.

💡Distribución

La distribución es la representación gráfica o numérica de los datos, mostrando la frecuencia de cada punto de datos. En el vídeo, se habla de distribuciones para entender cómo los puntajes se relacionan con percentiles y quartiles, y cómo se ven afectados por la media y la desviación estándar.

💡Media

La media es el promedio de un conjunto de datos y se menciona en el vídeo como la referencia para calcular el Z-score. Por ejemplo, si la media de una prueba es de 100, un puntaje de 130 tiene un Z-score de +2, lo que indica que está por encima de la media.

💡Estadísticas descriptivas

Las estadísticas descriptivas son medidas numéricas que resumen una colección de datos. En el vídeo, se explica cómo calcular diferentes estadísticas descriptivas en SPSS, como la media, la desviación estándar, los percentiles y los Z-scores, para entender la distribución de los datos.

💡Frecuencias

Las frecuencias son las estadísticas que muestran la cantidad de veces que cada punto de datos ocurre en una muestra. Aunque no se menciona directamente en el vídeo, están implícitas al hablar de percentiles y quartiles, ya que estos se calculan a partir de la distribución de frecuencias de los datos.

Highlights

Percentiles divide a set of observations into 100 equal parts.

The 50th percentile indicates that 50% of scores are below it.

A score of 33 at the 50th percentile means 50% of individuals scored below 33.

The 75th percentile shows 75% of scores are below it.

A score of 73 at the 75th percentile means 75% scored below and 25% scored above 73.

Quartiles divide data into four equal parts with values Q1, Q2, and Q3.

Quartile 1 is the 25th percentile, Q2 is the median (50th percentile), and Q3 is the 75th percentile.

A score of 33 equal to the 50th percentile is also the second quartile (Q2).

A score of 73 equal to the 75th percentile is the third quartile (Q3).

Percentiles and quartiles help understand the position of a score within a distribution.

Standard scores (z-scores) measure how many standard deviations a score is from the mean.

A z-score of +2 indicates a score is two standard deviations above the mean.

Z-scores help interpret scores in relation to the mean and standard deviation.

Scores with z-scores less than -1.96 or greater than 1.96 are considered unusual.

A score of 130 with a z-score of +2 is two standard deviations above the mean.

Z-scores can be calculated in SPSS using the 'Descriptives' function.

Descriptive statistics in SPSS can be calculated using 'Frequencies', 'Descriptives', and 'Explore' functions.

The 'Explore' function allows examining variables disaggregated by another variable.

Z-scores inform a score's position relative to the mean, different from mean and standard deviation which describe the distribution.

Transcripts

play00:00

[Music]

play00:09

welcome I'm Amanda rockinson-szapkiw

play00:30

going to briefly discuss how to

play00:31

calculate descriptive statistics in

play00:34

SPSS let's begin talking about

play00:37

specifically identifying defining and

play00:39

looking at examples of measures of

play00:42

position let's start with percentiles

play00:45

first now percen tiles can be defined as

play00:47

values that divide a set of observations

play00:49

into 100 equal parts or points in a

play00:52

distribution below which a given

play00:54

percentile or P of cases um lie let's

play00:59

look at an example of this let's say we

play01:00

have a distribution of scores and one of

play01:03

our scores is

play01:05

33 and let let's say this may be number

play01:09

of points that an individual scores on a

play01:11

test so we know that we have one

play01:13

individual in our distribution and they

play01:15

scored 33 points and we know that they

play01:20

their number of 33 is equal to the 50th

play01:23

percentile now if we look back at the

play01:26

definition of points in a distribution

play01:28

below which a given P of the cases lie

play01:32

we know that um if a score is it within

play01:36

the 50th percentile then 50% of the

play01:40

scores are below the 50th percentile so

play01:44

if we know this what then could we say

play01:46

about the score of

play01:53

33 well we could say that 50% of

play01:56

individuals in our distribution scored

play01:59

below

play02:00

33 let's take a look at another example

play02:04

and what percentiles are and how they're

play02:06

interpreted let's say that another

play02:09

person in our distribution scored a 73

play02:12

and we know that the score of 73 is

play02:14

equal to the 75th

play02:17

percentile so again looking back at our

play02:19

definition what we know is that 75% of

play02:22

scores are below the 70 75th percentile

play02:26

we could also look at it as 25% or above

play02:29

the 7

play02:30

5th

play02:31

percentile so knowing this what then can

play02:34

we say about the score of

play02:42

73 well what we could say is is that 75%

play02:46

of individuals scored below 73 and 25%

play02:50

of individuals scored above

play02:53

73 okay now that we have an

play02:56

understanding of percentiles let's move

play02:57

on to cor tiles cor tiles um are a rank

play03:02

rank order or they rank order the data

play03:04

into four equal parts the values that

play03:07

divide each part are called the first

play03:10

second and third cortile so they're

play03:13

denoted by q1 Q2 and Q3

play03:17

respectively so in terms of percentiles

play03:19

what we can say is that CTI that um

play03:22

quartiles can be defined as the as

play03:24

following for example quartile 1 is

play03:26

equal to the 25th percentile quartile 2

play03:29

2 is equal to the 50th

play03:32

percentile um or the median and quartile

play03:35

3 is equal to the 70th fth

play03:38

percentile so if we look at our score of

play03:41

33 and we know that 33 is with is is

play03:45

equal to the 50th percentile or the the

play03:48

median we can then

play03:51

say um that the score of 33 is equal to

play03:56

the second quartile or quartile 2 two do

play04:00

you see how how that that works so since

play04:03

we know that percent the 50th percentile

play04:05

is equal to the median the median is

play04:07

equal to the um second quartile and 33

play04:11

is equal to the 50th percentile we can

play04:14

say that 33 is equal to um quartile 2 or

play04:19

the second

play04:20

quartile then um let's look at 73 if we

play04:24

know the um score of 73 is equal to the

play04:27

705th percentile then what can we say

play04:30

it's quti what quartile is it equal

play04:33

to well as the 70th per 705th percentile

play04:38

is equal to the third quartile we can

play04:40

say that 73 equals the 3D

play04:43

quartile now that we have a basic

play04:46

understanding of percentiles and core

play04:48

tiles and as you can um see or as I hope

play04:51

you're seeing what these are helping us

play04:53

do is understand the specific position

play04:56

of a score so it's not necessarily

play04:59

helping us understand the distribution

play05:01

as we um as we as we looked at is at

play05:06

um oh as we looked at whenever we were

play05:09

looking at measures essential tendency

play05:11

in dispersion so measures of essential

play05:12

tendency and dispersion help us

play05:14

understand the overall distribution

play05:16

whereas percentiles and quartiles help

play05:18

us understand a specific score and how

play05:20

where it lies within that distribution

play05:22

so I'm hoping you're understanding that

play05:24

difference now before we move on let's

play05:26

go ahead and talk briefly about the

play05:29

standard scores um standard scores are

play05:32

how many standard deviations a score or

play05:36

an element is from the mean so standard

play05:38

scores are important for the reason of

play05:40

convenience and comparability of

play05:43

different data sets it makes it easier

play05:44

to interpret and probably the most

play05:47

common standard score used is the zcore

play05:50

and that's what we're going to focus on

play05:51

next so as I said a zcore is a

play05:54

standardized score here we see the

play05:56

formula for a zcore first we see the

play05:58

formula for a population and then we see

play06:00

one for a sample the formula for the

play06:03

sample is z is equal to x - M and X

play06:08

denotes the value or the number that

play06:10

you're looking at divided by the

play06:12

standard deviation narratively a zcore

play06:16

specifies the precise location of each

play06:20

value within a distribution in

play06:22

relationship to the mean um it gives us

play06:25

the number of standard deviations that a

play06:28

score lies either above or below the

play06:30

mean the sign of the zcore whether it's

play06:33

positive or negative then signifies

play06:35

whether the score is above the mean or

play06:37

below the

play06:39

mean so if a zcore is equal to zero that

play06:43

means that it is equal to the mean if a

play06:47

zcore is less than zero that means then

play06:51

it's less than the mean if a zcore is

play06:54

greater than zero then that means it's

play06:56

greater than the mean so if it if if we

play06:59

have a zcore of one that means that the

play07:02

number that we're looking at or the um

play07:04

value that we're looking at is one

play07:06

standard deviation greater than the mean

play07:09

whereas if we had a zcore negative to

play07:11

minus one let's say that would mean that

play07:13

the uh value that we're looking at is

play07:16

one standard deviation less than the

play07:19

mean if we had a zcore than of two we

play07:22

would say that the uh value that we're

play07:24

looking at is two standard deviations

play07:27

greater than the mean and so on and so

play07:31

forth let's take a look at zc scores a

play07:34

little bit more practically in light of

play07:36

an example

play07:37

scenario um let's consider a student's

play07:40

sample population of normally

play07:42

distributed educational Statistics final

play07:45

exam scores that have a mean of 100 and

play07:48

a standard deviation of

play07:50

15 if we know this what can we say about

play07:54

an individual who scores

play07:58

130 well well we can say something about

play08:01

that score if we know its zcore and

play08:04

remember to calculate the

play08:06

zcore what we will um do is we will take

play08:10

x minus the mean so 130 minus the mean

play08:15

of 100 divided by the standard deviation

play08:17

of 15 and what we find out is is that

play08:21

the score of 130 has a zcore of +

play08:26

2 so what does this zcore of plus 2 mean

play08:30

well based on what we just talked about

play08:32

the score is two standard deviations

play08:35

above the mean now here it's important

play08:37

to note by convention outcomes that have

play08:40

zcore values less than minus 1.96 or

play08:45

greater than 1.96 are usually viewed as

play08:49

unusual or extreme in a normal

play08:51

distribution what we know is is that

play08:53

about

play08:54

2.5% of the area lies below um the zcore

play09:00

of minus 1.96 and

play09:03

2.5% of the area lies above the zcore of

play09:07

plus uh

play09:09

1.96 so together these two taals make

play09:13

the most extreme 5% of the outcomes in

play09:16

the distribution of scores that's why we

play09:18

would consider scores above uh 1

play09:22

196 unusual as well as scores below up

play09:26

1.96

play09:28

unusual so going back to our example

play09:32

what can we say about our score of 130

play09:36

that has a zcore of

play09:39

two well based on convention can we not

play09:42

say that this the person that scored 130

play09:46

on the educational statistics file or

play09:48

final was extremely

play09:56

high let's consider extreme or un usual

play09:59

Z scores in light of another example

play10:02

let's consider extreme Z scores in terms

play10:04

of

play10:05

height we know that women's height in

play10:08

the US is approximately normally

play10:10

distributed with a mean of 64 in and a

play10:13

standard deviation of 2.56 in now I want

play10:17

you to consider that you walk into a

play10:19

room of women and what you see is a

play10:23

woman who is 67 in tall or 66 in tall

play10:27

and a woman who is is 71 in tall okay so

play10:32

you see these two women which woman are

play10:35

you more likely to take note of the one

play10:38

that's 66 in or the one that's 71

play10:42

in you probably said the one that's 71

play10:46

in because we know that the average

play10:49

height of a woman is 64 in so most of

play10:51

the women in the room are going to be

play10:53

around 64 in and 66 is close to 64

play10:58

however 7 one's not so close and chances

play11:01

are the woman who's 71 in is going to

play11:03

look a lot taller than everyone else in

play11:06

the room in fact her height may be

play11:09

considered unusual or

play11:12

extreme now let's look at these two

play11:14

women's Z scores um and see if the if

play11:18

the conclusion we just made holds true

play11:21

here we'll see that the woman that was

play11:23

66 in if we calculate her zcore by

play11:27

taking her uh um her inches minus her or

play11:33

minus the mean divided by the standard

play11:35

deviation that her zcore is

play11:38

1.17 and what we note by convention is

play11:41

this is an unusual or extreme we can

play11:44

then calculate the a zcore for the woman

play11:47

that's 71 in and what we note here is

play11:50

that her zcore is

play11:53

2.73 and by convention that's above a

play11:57

1.96 and therefore or her height could

play12:00

be considered

play12:02

extreme now understanding Z scores can

play12:06

be helpful in many ways because these

play12:07

scores really help us understand a

play12:10

person's position in a

play12:12

distribution we just looked at the

play12:14

example of height but let's look at how

play12:16

a zcore may be useful if we go back to

play12:18

our educational example an educator may

play12:21

want to know again um may want to know

play12:24

or better understands let's say a

play12:27

specific students achievement score on

play12:29

multiple assignments in a

play12:31

course um or on a specific assignment in

play12:34

a course examining Z scores for a

play12:37

student on each assignment or multiple

play12:40

assignments can inform the educator if a

play12:42

specific student is performing average

play12:45

or extremely high or extremely low and

play12:48

this can then inform their instruction

play12:50

so again um a zcore helps inform the

play12:54

position of one's score which is a

play12:56

little different than standard deviation

play12:58

and mean which tells us about the entire

play13:01

distribution now that we understand um

play13:06

zores let's move on and talk a little

play13:08

bit about how to calculate both them and

play13:10

other descriptive statistics in

play13:13

SPSS descriptive statistics can be

play13:16

calculated using three different

play13:19

functions or primarily three different

play13:20

functions in SPSS and that's frequencies

play13:22

descriptives and explore now there's

play13:25

some overlap in these functions in that

play13:28

they calculate mean and standard

play13:30

deviation however they do have different

play13:34

features for example frequency allows us

play13:36

to calculate specific percentiles

play13:39

whereas the explore function only allow

play13:41

or only calculates uh percentiles preset

play13:45

by the SPSS software and the frequency

play13:48

function also allows you to um look at

play13:51

one variable at a time however the

play13:53

explore function enables you to examine

play13:56

a variable disaggregated by another

play13:58

variable so if I wanted to look at let's

play14:00

say course points disaggregated by

play14:02

gender or ethnicity I would use the

play14:04

explore function whereas if I was just

play14:07

interested in examining course points I

play14:10

might use the frequency function so I

play14:13

encourage you to explore the different

play14:15

functions um in which you can calculate

play14:17

the descriptive statistics in SPSS and

play14:20

again those functions are frequency

play14:21

descriptives and

play14:23

explore if you're interested in

play14:26

calculating a zcore then you'll want to

play14:28

use the descriptives function so you'll

play14:31

go to analyze descriptive statistics and

play14:33

click descriptives once you've done that

play14:36

you'll choose what variable you want to

play14:38

analyze and below the variable list what

play14:40

you'll see is a little button that you

play14:43

can tick that says save standardized

play14:45

values as variables if you desire to

play14:48

calculate the zores for your entire

play14:51

sample population you tick this and then

play14:53

you tick okay so that's how you

play14:56

calculate Z scores in SPSS

play15:00

this now concludes part two of our

play15:01

tutorial on descriptive statistics you

play15:04

should now be able to identify and

play15:06

provide examples of different measures

play15:08

of phys including percentiles quartiles

play15:11

and zores and you should also understand

play15:14

how to calculate descriptive statistics

play15:16

and specifically zc scores in SPSS

Rate This

5.0 / 5 (0 votes)

Étiquettes Connexes
EstadísticasSPSSPercentilesCuartilesZ-ScoreAnálisis de datosTutorialEducaciónEstadísticas descriptivasDistribución de datos
Besoin d'un résumé en anglais ?