de.mpg.escidoc.pubman.appbase.FacesBean

Post

 
 Vis
  Evidensbaseret arbejde på det sociale område i teori og praksis
Item is

Ophav

 Ophav:
Hawgood, Anna Esther1, Forfatter
Ploug, Niels2, Vejleder
Tilknytninger:
1Det Samfundsvidenskabelige Fakultet, Københavns Universitet, København, Danmark, diskurs:7001              
2Økonomisk Institut, Det Samfundsvidenskabelige Fakultet, Københavns Universitet, København, Danmark, diskurs:7014              
skjul Ophav
Vis Ophav

Indhold

Ukontrollerede emneord: evidensbaseret arbejde, Campbell, effektmåling, socialt arbejde
 Abstract: The aim of this thesis is to analyse to which degree it is possible for the Danish Social Services to produce and implement credible evidence by carrying out research locally.
In order to answer this question, I have set up an ideal model for conducting research concerning the effect of social initiatives, using Campbell’s Causal Model. I then analysed to which extent the ideal model can be implemented in the Danish Social Services, using the child protective services in Copenhagen as an example. The focus of this analysis is on how the choice of research design affects the validity of the results. In my discussion I have examined whether it is possible to identify an optimal research design for measuring the effect of social initiatives, and to which degree the quality of a study can be assessed on research design alone. I have then discussed in which way the locally conducted research fits into the paradigm of evidence-based social work, and how the research results may be implemented in the social services to ensure that they lead to evidence-based practices among social workers.
I have set up the ideal model by first defining what constitutes a credible study on the effect of social initiatives, and then examining the degree to which different research designs achieve credibility of results, using the methodology set up in Campbell’s Causal Model. In this methodology, credibility of a study follows from validity of the results. In order to secure the credibility of a study, it is therefore necessary to minimize the plausibility of all threats towards validity. I have illustrated the ideal model for conducting research concerning effect of social initiatives by comparing the ability of four research designs’ ability to minimize the threats towards validity.
The results show that none of the research designs are capable of excluding all of the threats towards validity. The final degree of plausibility will only vary with choice of research design for 12 of the threats towards validity when corrective measures have been implemented. These 12 threats mainly affect internal validity. The results show that randomized control trials are most suitable for exclusion of threats towards internal validity. The analysis also shows that, despite randomized control trials are most suitable for exclusion of threats towards internal validity, it is not unequivocally clear which research design will be most suitable for achieving credible results in studies concerning effect of social initiatives.
Finally, the analysis shows that several of the measures to minimize threats towards validity must be implemented at specific points in the study. The structure of the research process is therefore as important for the validity of the results, as is the choice of research design.
Since the variation between research designs concerning the degree to which the threats’ plausibility can be minimized is so low, the choice of research design must depend not only on the degree to which different research designs achieve validity, but also on which types of validity and threats to validity are deemed central to the question of credibility, the resource consumption associated with each research design, and whether it is realistic to conduct a study with the chosen research design.
If a study is well structured, it will become more apparent which threats to validity will critically affect the results, and therefore also which research design will be most suitable for the study. If the locally conducted research is well structured, and the choice of research design is based on Campbell’s Causal Model, it will produce credible evidence which can be used to increase the knowledge base of the practitioners. The role of the locally conducted research will not be to replace, but rather to supplement, the evidence produced by SFI Campbell and similar entities.
skjul Indhold
Vis Indhold

Filer

Navn:
Endeligt(1).pdf (Hovedtekst)
Bemærkninger:
-
Tilgængelighed:
Offentlig
Mime-type / størrelse:
application/pdf / 993KB
Copyright dato:
2013-10-04
Copyright information:
De fulde rettigheder til dette materiale tilhører forfatteren.
skjul Filer
Vis Filer

Basal

Bogmærk denne post: https://diskurs.kb.dk/item/diskurs:52902:1
 Type: Speciale
Alternativ titel: Evidence-based social work in theory and practice
skjul Basal
Vis Basal

Links

Vis Links

Detaljer

Sprog: Danish - dan
 Datoer: 2013-08-26
 Sider: -
 Publiceringsinfo: København : Københavns Universitet
 Indholdsfortegnelse: Summary 1
Forord . 4
1. Indledning . 7
1.1 Formål og problemformulering 7
1.2 Afgrænsning . 8
Figur 1 – Udvælgelse af individer til evalueringer 10
2. Baggrund 11
2.1 Evaluering på det sociale område . 11
2.2 Debatten omkring evidensbaseret arbejde på det sociale område 13
2.3 Forandringskompasset 14
3. Metode 16
Figur 2 – Proces for plausibilitetsanalyse 17
3. Analyse . 17
3.1 Validitetstyper 18
3.1.1 Statistisk validitet 18
3.1.2 Intern validitet . 20
3.1.3 Begrebsvaliditet . 22
3.1.4 Ekstern validitet . 26
3.2 Forskningsdesign 28
3.2.1 Design 1 - Før- og eftermålinger på én gruppe . 29
3.2.1.1 Plausibilitetsanalyse for validitetstrusler i Design 1 29
3.2.2 Design 2 - Forsøg med ikke-ækvivalente kontrolgrupper . 34
3.2.2.1 Plausibilitetsanalyse for validitetstrusler i Design 2 34
3.2.3 Design 3 - Gentagen institutionel cyklus design . 38
3.2.3.1 Plausibilitetsanalyse for validitetstrusler i Design 3 38
3.2.4 Design 4 - Randomiserede kontrollerede forsøg . 42
3.2.4.1 Plausibilitetsanalyse for validitetstrusler i Design 4 43
3.2.5 Plausibilitetsanalyse for generelle validitetstrusler . 45
3.3 Opsummering af hovedresultater . 52
Tabel 2 – Optælling af validitetstrusler indenfor hver udelukkelseskategori 53
Tabel 1 – Validitetstrusler for Design 1 til 4 54
Tabel 3 – Procesplan for minimering af truslers plausibilitet 55
4. Diskussion 56
4.1 Forskningsbaserede evalueringer i de sociale myndigheder 56
4.1.1 Prioritering af validitetstyper 56
4.1.2 Strukturering af evalueringsprocessen 57
4.1.3 Kan de foreslåede korrigerende foranstaltninger gennemføres? . 59
4.1.4 Relevanskriteriet 60
4.1.5 Valg af forskningsdesign . 61
4.2 Campbells Kausale Model i lys af debatten om pålidelig evidens . 65
4.3 Velfærdssystemers betydning for valg af forskningsdesign 67
4.4 Implementering af evidensbaseret arbejde på det sociale område . 69
5. Konklusion . 73
6. Perspektivering . 74
Referencer 75
 Note: -
 Type: Speciale
skjul Detaljer
Vis Detaljer

Kilde

Vis Kilde